Research ArticleDevelopment of Estimation Force FeedbackTorque Control Algorithm for Driver Steering Feel in VehicleSteer by Wire System Hardware in the Loop
Sheikh Muhammad Hafiz Fahami1 Hairi Zamzuri2 and Saiful Amri Mazlan2
1UTM-Razak School of Engineering and Advanced Technology Universiti Teknologi Malaysia Jalan Semarak54100 Kuala Lumpur Malaysia2Vehicle System Engineering I-Kohza North Wing MJIIT Universiti Teknologi Malaysia Jalan Semarak54100 Kuala Lumpur Malaysia
Correspondence should be addressed to Hairi Zamzuri hairiklutmmy
Received 11 November 2014 Revised 10 April 2015 Accepted 30 April 2015
Academic Editor C S Shankar Ram
Copyright copy 2015 Sheikh Muhammad Hafiz Fahami et al This is an open access article distributed under the Creative CommonsAttribution License which permits unrestricted use distribution and reproduction in any medium provided the original work isproperly cited
In conventional steering system a feedback torque is produced from the contact between tire and road surface and its flows throughmechanical column shaft directly to driverThis allows the driver to sense the steering feel during driving However in steer by wire(SBW) system the elimination of the mechanical column shaft requires the system to generate the feedback torque which shouldproduce similar performance with conventional steering system Therefore this paper proposes a control algorithm to create theforce feedback torque for SBW system The direct current measurement approach is used to estimate torque at the steering wheeland front axle motor as elements to the feedback torque while adding the compensation torque for a realistic feedback torqueThegain scheduling with a linear quadratic regulator controller is used to control the feedback torque and to vary a steering feel gainTo investigate the effectiveness of the proposed algorithm a real-time hardware in the loop (HIL) methodology is developed usingMatlab XPC target toolbox The results show that the proposed algorithm is able to generate the feedback torque similar to EPSsteering system Furthermore the compensation torque is able to improve the steering feel and stabilize the system
1 Introduction
The latest generation of integrated steering system is steer bywire (SBW) systemwhich is also known as independent steer-ing system It eliminates the need of mechanical column shaftbetween the steering wheel and the front axle system and wasreplaced with sensor actuators and electronic controller unit(ECU) as shown in Figure 1 The SBW system offers numer-ous advantages such as large space in cabin [1ndash3] While forvehicle manufacturers the absence of mechanical columnshaft gives maximum freedom interior design and ergonom-ics thus making it more comfortable for the driver [4 5]Moreover it reduces the driver impact force from the frontalaccident The SBW system is also potential to enhance thevehicle performance such asmaneuverability and stability [1]
There are several characteristics of SBW system [1] Oneof the characteristics of a SBW system is feedback torquefor driver steering feel whereby it is the most challengingissue in SBW system [6 7] The function of feedback torqueis for driver steering feel and steering wheel returnabilitywhen drivers release their hand from the steering wheel [8]Previous researchers have done several studies on how tocreate the feedback torque Based on control parameter ofvehicle speed and steeringwheel angleOh et al [9] developeda torque map to generate the feedback torque The generatedtorque quantity is small at low speed and increased at highspeed Authors proposed to control a gain for each controlparameter in order to vary the feedback torque Authors alsoclaimed that the torque map could give engineers freedomto tune for a real driver steering feel Kim et al proposed that
Hindawi Publishing CorporationInternational Journal of Vehicular TechnologyVolume 2015 Article ID 314597 17 pageshttpdxdoiorg1011552015314597
2 International Journal of Vehicular Technology
Steering wheel system
Steering wheel
DC motor
Positionsensorencoder
ECU
Electronic controller unit Wiring
Front axle system
DC motorPositionsensorencoder
Rack and pinion Tire
Steer by wire (SBW) system
Steering wheel
CouplingMechanical column shaft
Conventional steering system
Figure 1 Difference between SBW system and conventional steering system
improvement to the torquemap has beenmade by adding thedamping torque to create more realistic driver steering feeland to improve steering wheel returnability [10] MeanwhileAmberkare et al [6] used the steering wheel angle and fed tomodel transfer function to create a feedback torque By usingthe appropriate transfer function it is possible to changethe desired steering feel by varying a proper gain of thetransfer function The approach of the disturbance observeris introduced by Asai et al [11] The position and current ofthe front axle motor are used to generate the feedback torqueFurthermore a reference model that consists of the inertiaand damping factor of steering wheel motor is proposed byPark et al [12] to generate the feedback torque The modelmatching technique which combines the mechanical andelectrical parts has been introduced by Odenhtal et al [13]to generate the feedback torque while the series of vehiclecornering control algorithm is taken into consideration tooptimize the feedback torque as proposed by Na et al [14]
The main objective of this paper is to proposes a controlalgorithm to generate the force feedback torque for driversteering feel in the vehicle SBW system The subsystem ofSBW system which is steering wheel and front axle system ismodelled and the control algorithm is explained To verify theeffectiveness of the proposed algorithm the hardware in theloop (HIL) is presented and interfaced using a Matlab XPCtarget toolbox The result was then is compared with electricpower steering (EPS) system [15]
2 The Modelling of Steer byWire (SBW) System
The SBW system is divided into two subsystems that consistof steering wheel and front axle system The system diagramof subsystem is shown in Figure 2
Table 1 Steering wheel system parameter
Parameter Description Value Unit1198771 Motor resistance 564 ohm1198711 Motor inductance 0017 henry119870119904119898
Motor constant 0024 Nm1198691198981 Motor inertia 00036 kgm2
1198871198981 Motor damping 00068 Nm(rads)1198961199041 Lumped torque stiffness 0025 Nm
1198811199041 Voltage source mdash Volt
21 Steering Wheel Model The main purpose of steeringwheel motor is to generate the steering feedback torque fordriver steering feel Figure 2(a) shows the system diagramfor the steering wheel system The input to the system is thetotal feedback torque (120591total feedback) and the output is the ratechange of steering wheel motor angular displacement ( 120575
1198981)the motor angular displacement (120575
1198981) and the current ofsteering wheel motor (119894
1198861) The parameters of the steeringwheel system are illustrated in Table 1
The mathematical equations of the steering wheel systemare written as follows
The steering motor angular displacement is
1205751198981 = minus
11988711989811198691198981
1205751198981 +
1198961199041
1198691198981
1198941198861 (1)
and the steering motor current is
1198941198861 = minus
11987711198711
1198941198861 +
119896119904119898
1198711
1205751198981 +
1198811199041
1198711 (2)
International Journal of Vehicular Technology 3
120591totalfeedback
ks1
120575sw
bm1
120575m1
Jm1
L1
R1
Vs1
ia1
(a)
Vs2
L2
R2
ia2
bm2
120575m2
Jm2
ks2
brackf
Mrackf
yrackfklf klfBkpf Bkpf
Jm2 Jm2
(b)
Figure 2 System diagram (a) steering wheel and (b) front axle system [15]
Consequently the state equation of steering wheel system isgiven as
119904119898
= 119860119904119898
119909119904119898
+119861119904119898
119906119904119898
119910119904119898
= 119862119904119898
119909119904119898
119909119904119898
= [ 1205751198981 1205751198981 1198941198861]
(3)
And 119906119904119898
is considered the input of the steering wheel motorand the parameter states are
119860119904119898
=
[[[[[[
[
minus(11988711989811198691198981
) 01198961199041
1198691198981
1 0 0
minus(119896119904119898
1198711) 0 minus(
11987711198711
)
]]]]]]
]
119861119904119898
=
[[[[
[
0
0
11198711
]]]]
]
119906119904119898
= 120591total feedback = 1198811199041
(4)
22 Front Wheel System The function of the front axle sys-tem is to ensure the front tire angle follows the steering wheelangle command according to the steering ratio [17]The frontaxle is composed of a front axle DC motor rack and pinionand tire model The system diagram of the front axle systemand the subsystem block diagram are shown in Figures 2(b)and 3 respectively The input to the system is the steeringwheel angle (120575
119904119908) while the front tire angle (120575
119891) is considered
the output The parameters of the front axle system areapproximated assumption that are illustrated in Table 2
Table 2 Front axle system parameter [15]
Parameter Description Value Unit1198772 Motor resistance 464 ohm1198712 Motor inductance 0015 henry119870119891119898
Motor constant 0032 Nm1198691198982 Motor inertia 00062 kgm2
1198871198982 Motor damping 00036 Nm(rads)1198961199042 Lumped torque stiffness 0025 Nm
119861rack Rack damping coefficient 0015 henry119872rack Rack lumped mass 0032 Nm119896119897119891
Rack linkage stiffness 000062 kgm2
119903119871
Offset of king pin axis 000036 m119903119901
Pinion gear radius 0025 m119861119896119901
King pin damping coefficient 000062 kgm2
119868119891
Lumped front wheel inertia 000036 kgm2
1198961199042 Lumped torque stiffness 0025 Nm
120575sw 120575m2 120575fyrack
Front axle system
Front axleDC motor
Rack andpinionsystem
Front tiredynamics
Figure 3 Block diagram of front axle subsystem
The mathematical equation of the front axle DC motor iswritten as follows
The front axle motor angular displacement is
1205751198982 = minus
11988711989821198691198982
1205751198982 +
1198961199042
1198691198982
1198941198862 (5)
4 International Journal of Vehicular Technology
and the front axle motor current is
1198941198862 = minus
11987721198712
1198941198862 +
119896119891119898
1198712
1205751198982 +
1198811199042
1198712 (6)
Consequently the state equations of the front axle motor aregiven as
119891119898
= 119860119891119898
119909119891119898
+119861119891119898
119906119891119898
119910119891119898
= 119862119891119898
119909119891119898
119909119891119898
= [ 1205751198982 1205751198982 1198941198862]
(7)
And 119906119891119898
is the input to the front axle motor while outputis the rate change of front axle motor angular displacement( 1205751198982) the front axle motor angular displacement (120575
1198982) andthe current of front axle motor (119894
1198862)The parameter states are written as follows
119860119891119898
=
[[[[[[[
[
minus(11988711989821198691198982
) 01198961199042
1198691198982
1 0 0
minus(
119896119891119898
1198712) 0 minus(
11987721198712
)
]]]]]]]
]
119861119891119898
=
[[[[
[
0
0
11198712
]]]]
]
119906119891119898
= 120575119904119908
= 1198811199042
(8)
The rack and pinion system is
119910rack =1
119872rack[(minus
2119896119897119891119910rack
1199032119871
)minus(1198961199042119910rack1199032119901
)
minus119861rack119891 119910rack +(11989611990421205751198982119903119901
)]
(9)
And 119906119910rack is considered the input of the rack and pinion
system and the output is the rack displacement (119910rack) andthe parameter states are as follows
119860119910rack
=[[[
[
0 1
minus[(
2119896119897119891
119872rack1199032119871
) + (1198961199042
119872rack1199032119901
)] minus(119861rack119872rack
)
]]]
]
119861119910rack =
[[
[
0
(1198961199042
119872rack119903119901)
]]
]
119906119910rack = 120575
1198982
(10)
Fyr
lr lf
r
120572r
120573
Fyf120572f
120575f
Figure 4 Single track-linear vehicle model
and the front tire dynamic is
120575119891=
1119868119891
[(minus119896119897119891120575119891+(
119896119897119891119910rack
119903119871
)minus119861119896119901
120575119891)] (11)
and 119906tire is considered the input of the front tire dynamicTheoutput is the front tire angle (120575
119891) and the parameter state is
119860 tire =[[
[
0 1
minus(
119896119897119891
119868119891
) minus(
119861119896119901
119868119891
)
]]
]
119861tire =[[
[
0
(
119896119897119891
119868119891119903119871
)
]]
]
119906tire = 119910rack
(12)
23 Single Track-Linear Vehicle Model The stability charac-teristics of a steer-by-wire vehicle are affected by the influenceof front tires via the feedback of the vehicle dynamic responseThus it is necessary to develop a single track-linear vehiclemodel as mentioned by author [16 18ndash21] in simulation andexperimental validationThe single track linear vehiclemodelis shown in Figure 4 In this paper the aim is to model thelinear vehicle in order to generate the estimation of self align-ing torque (120591
119886) This acts as the disturbance input to the front
axle system and assures the road contact with front tire Thedynamics of single track-linear vehicle model is representedby bicycle model where the states contain yaw rate (119903) andvehicle body slip angle (120573)
The following assumptions are considered for normaldriving maneuverer These assumptions could simplify thevehicle model [18]
(i) friction force in the 119909-direction is negligible
(ii) vehicle speed is constant
(iii) angle for right and left tire is approximately the same
International Journal of Vehicular Technology 5
With the aforementioned assumptions the derived equationsof the motion of yaw rate (119903) and vehicle body slip angle (120573)are shown in (13) as follows
119903 = [
[
(1198972119891119862119891+ 119897
2119903119862119903)
119868119904119881
]
]
119903+[
[
(1198972119903119862119903minus 119897
2119891119862119891)
119868119904
]
]
120573
+[
(119897119891119862119891)
119868119904
]120575119891
120573 = [1+(119897119903119862119903minus 119897119891119862119891)
1198981198812 ] 119903 minus[
(119862119891+ 119862119903)
119898119881]120573
+[
119862119891
119898119881]120575119891
(13)
In the linear tire region lateral force at the front and rear isrelated to slip angle by cornering stiffness of the front and reartires written as follows
119865119910119891
= minus119862119891120572119891
119865119910119903
= minus119862119903120572119903
(14)
At a larger slip angle the relation between lateral force andslip angle becomes nonlinear By assuming a smaller slipangle approximation the slip angle is written in terms ofvehicle body slip angle and yaw rate that are considefed asfollows
120572119891= (120573+
119897119891
119881119903minus 120575119891)
120572119903= (120573+
119897119903
119881119903minus 120575119903)
(15)
Rewriting (13) into state space form we haveV119898 = 119860V119898119909V119898 +119861V119898119906V119898
119910V119898 = 119862V119898119909V119898
119909V119898 = [119903 120573]
(16)
And 119906V119898 is considered the input to the linear vehicle modelwhile the yaw rate (119903) and vehicle body slip angle (120573) arethe output of the linear vehicle model The parameter stateis written as follows119860V119898
=
[[[[[[
[
(
(1198712119891119862119891+ 119871
2119903119862119903)
119868119904119881
) minus
(119871119903119862119903minus 119871119891119862119891)
119868119904
1 + (
(119871119903119862119903minus 119871119891119862119891)
1198981198812 ) minus
(119862119891+ 119862119903)
119898119881
]]]]]]
]
119861V119898 =[[[
[
119871119891119862119891
119868119904
119862119891
119898119881
]]]
]
119906V119898 = 120575119891
(17)
The parameter of a linear vehicle model directly can bemeasured but some parameters need to be estimated Thereare several methods to estimate themodel parameter [22 23]A method based on Sahoo et al [22] is used to estimate themodel parameters with assumption that the vehicle mass isdistributed to front and rear of each tire The total of vehiclemass (119898) that is based on the front (119898
119891) and rear (119898
119903) tire
masses is written as follows
119898 = 119898119891+119898119903
(18)
when
119898119891= 119898119891119897+119898119891119903
119898119903= 119898119903119897+119898119903119903
(19)
where 119898119891119897 119898119891119903 119898119903119897 and 119898
119903119903are the masses of the vehicle
at front left front right rear left and rear right respectivelyThe parameter length of wheel base (119897) which is sum of lengthfrom a center of front tire 119897
119891and rear tire 119897
119903to COG of vehicle
body is written as follows
119897 = 119897119891+ 119897119903 (20)
while the inertia of vehicle 119868119904that is approximated by assum-
ing the vehicle as two point masses joined with mass-lessroad [22] is written as follows
119868119904= 1198981198911198972119891+1198981199031198972119903 (21)
The front cornering stiffness 119862119891and rear cornering stiffness
119862119903parameters are written in (22) by assuming that the type
of radial tire is used and cornering stiffness per degree of slipangle is approximately 16-17 of the load on the tire [22]Consider
119862119891= 119898119891lowast 119892 lowast 0165 (Ndeg)
119862119903= 119898119903lowast 119892 lowast 0165 (Ndeg)
(22)
The parameters of the linear vehicle model are illustrated inTable 3
3 Force Feedback Torque EstimationControl Algorithm
It has been known that the force feedback torque for driversteering feel should be created artificially due to the elimina-tion of mechanical column shaft in SBW system
Moreover the proportion of the steering feel obtainedshould be similar to conventional steering system For thisreason the steering wheel in SBW system that is equippedwith the motor actuator is used to generate the steeringfeel by controlling the total feedback torque (120591total feedback)A schematic block diagram of the proposed force feedbacktorque control is shown in Figure 5
Based on Figure 5 the total force feedback torque consistsof the torque of front axle motor (120591
119891119898) torque on steering
wheel motor (120591119904119898) and compensation torque which is the
inertia (120591inertia) and damping torque (120591damp) Meanwhile the
6 International Journal of Vehicular Technology
+
+
+
minus
minus
Wheel synchronization
120575sw
120575sw
120591a
Frontaxle
system
LQRcontroller
LQRcontroller
120575m2
120575m2
120575f
ia2
120573
r
Linearvehiclemodel
Force feedback torque
120591map
120591fm
120591sm
120591damp
120591inertia
VGain
scheduling
kfm
kfeelsum
120591total feedback Steeringwheelsystem
Steering wheel system
Steeringwheel
120575m1
120575m1
ia1
ksm
Estimation selfaligning torque
Torque(RMS)
Figure 5 Overview of force feedback torque control algorithm for SBW system
Table 3 Linear vehicle model [16]
Parameter Description Value Unit119862119891 Front cornering stiffness 40000 Nrad
119862119903 Rear cornering stiffness 35000 Nrad
119897119891
Length from center frontwheel to (COG) vehicle 14 m
119897119903
Length from center rearwheel to COG vehicle 1 m
119898 Mass of the vehicle 1535 kg119868119904 Inertia of vehicle 2149 kgm2
119881 Speed of vehicle mdash kmh119865119910119891 Front tire lateral force mdash Nm
119865119910119903 Rear tire lateral force mdash Nm
120572119891 Front tire slip angle mdash Nm
120572119903 Rear tire slip angle mdash Nm
torque map is based on the function of the steering wheelangle and vehicle speed is added to the total feedback torqueMoreover a steering feel gain (119896feel) is adapted based on thefunction of steering wheel angle and vehicle speed to varythe feedback torque In order to control the feedback torquethe gain scheduling with LQR controller is used To showthe influence of the tire on the road surface the self aligningtorque (120591
119886) acts as input disturbance to the front axle system
as follows
120591119886= minus119862
119891(119905119901+ 119905119898) [120573+
119897119891119903
119881minus 120575119891]120583 (23)
where 120583 is coefficient for road dry condition pneumatic trail(119905119901) and mechanical trail (119905
119898)
On the other hand any change of the road surface couldchange the feedback torque magnitude of the front axlemotor Furthermore the LQR controller of the front axle sys-tem is used to synchronize the front tire angle (120575
119891) with driver
input steering wheel angle (120575119904119908) by controlling the angle of
the front axle motor (1205751198982) in accordance with the steering
ratioThus based on the proposed control algorithm the total
force feedback torque to the steering wheel system is writtenas follows
120591total feedback = ([(radicmean (120591119904119898
+ 120591119891119898
+ 120591map))]
+ 120591inertia + 120591damp) 119896feel
(24)
Author [9 10] has generated and discussed the force feed-back torque using the torque map conceptThe torque map isbased on the function of the steering wheel angle and vehiclespeed parameter The feedback torque is determined by twocontrol parameters (119896
120573) for the vehicle speed and (119896
120572) for the
steering wheel angle functionThe torque based on the function of vehicle speed is given
as follows [9]
1205911 = 1198961205731199092(13119909minus
12119881max)+119879in (25)
International Journal of Vehicular Technology 7
0 2 4 6 8 10minus40
minus30
minus20
minus10
0
10
20
30
40
Time (s)
Stee
ring
torq
ue (N
m)
70kmh75kmh
50kmh40kmh30kmh20kmh10kmh
80kmh
60kmh
Figure 6 Steering torque based on torque map at different speed
where (119909) is initial velocity and (119881max) is maximum velocitywhile the torque based on a function of steering wheel angleis written as follows [9]
1205912 = 119896120572radic1003816100381610038161003816120575119904119908
1003816100381610038161003816 sgn 120575119904119908 (26)
Based on the torque map concept the results of steeringtorque at different vehicle speed are shown in Figure 6 whenthe input steering wheel angle is a double lane changemanoeuvreThe torque increased at high speed anddecreasedat low speed [9 24] By varying the two control parametersthe steering reactive torque changes at different torque mag-nitude However the feedback torque based on the torquemap concept did not create the artificial steering feel dueto the fact that no element of road contact such as vehicledynamic responses are feedback to the torque map modelThus driver not sense realistic driver steering feel
Thus for this reason the front axle motor torque andsteering wheel motor torque are added to the torque mapThese ensure that the driver senses the feedback torque fromroad contact The front axle motor torque is chosen due toits proportion to the road surface The reaction torque of theroad surface gives the effect to the rack and pinion systemand it connects to the front axle DC motor Thus the frontaxle motor torque would always be proportional to the roadsurface Any changes in road surface will influence the frontaxle motor torque In order to measure the torque at the frontaxle motor the direct currentmeasurement technique is usedto estimate the torque
The direct relationship between torque and current of amotor can be defined [25]Thus then front axlemotor torque(120591119891119898
) that consists of the front axle motor constant (119896119891119898
) andthe motor current (119894
1198862) is written as follows
120591119891119898
= 1198941198862119896119891119898 (27)
and the corresponding steeringwheelmotor torque (120591119904119898) that
consists of the steering wheel motor constant (119896119904119898) and the
motor current (1198941198861) is written as follows
120591119904119898
= 1198941198861119896119904119898 (28)
0 2 4 6 8 10Time (s)
Mot
or to
rque
(Nm
)
Front axle motor torque
Steering wheel motor torque6
4
2
0
minus2
minus4
minus6
Figure 7 Steering wheel and front axle motor torque at 80 kmh
Figure 7 shows the front axle motor and steering wheelmotor torque response when the input is a double lanechangemanoeuvreThe results show that the front axlemotorgives a high torque This is due to the fact that the inputdisturbance torque between the tire and road surface which isthe self aligning torque has an effect on the front axle motorIncreasing the vehicle speedwill increase the front axlemotortorque On the other hand the magnitude of steering wheelmotor torque is decreased due to the elimination of mechan-ical column shaft and is not affected by the self aligningtorque
The phase compensation torque consists of the inertiatorque and damping torque potential to recreate the artificialsteering feel of the conventional steering system Moreoverit is intended to adjust the steering feel reduce the vibrationand stabilize the system Due to this fact the element of com-pensation torque is taken in order to improve the feedbacktorque for a driver steering feel and to stabilize the system
The damping torque (120591damp) which is the rate change ofthe steering wheel motor angle ( 120575
1198981) and the damping gain(119896damp) is written as follows
120591damp = 1205751198981119896damp (29)
The inertia torque (120591inertia) that is the acceleration of thesteering wheel motor angle ( 120575
1198981) and the inertia gain (119896inertia)is written as follows
120591inertia = 1205751198981119896inertia (30)
The inertia and damping torque responses are shown inFigure 8 when the input steering wheel angle is a double lanechange manoeuvre It is obtained that the damping torquegives a smooth signal and reduces the vibration Meanwhilethe inertia torque gives high magnitude Thus the combina-tion between the inertia torque and damping torque not onlyprovides an average of steering feel but is also able to stabilizethe system at the same time Figure 9 shows the feedback
8 International Journal of Vehicular Technology
0 2 4 6 8 10minus6
minus4
minus2
0
2
4
6
Time (s)
Torq
ue (N
m)
Inertia torque Damping torque
Figure 8 Inertia and damping torque
0 2 4 6 8 10Time (s)
Forc
e fee
dbac
k to
rque
(Nm
)
15
10
5
0
minus5
minus10
minus15
minus20
minus25
10kmh
50kmh100 kmh120 kmh
135kmh
Figure 9 Proposed force feedback torque response at differentvehicle speed
torque response based on (24) The torque is increasedaccording to the vehicle speed The higher vehicle speed willincrease the feedback torque and vice versa due to the samefact that is proposed by torque map
31 Linear Quadratic Regulator (LQR) Controller The LQRcontroller provides the best possible performance to controland stabilize the system by changing the location of the polesof the system to the optimal location for a time responseovershoot and steady stateThis is done by designing the statefeedback control119870 thatminimized cost function 119869 to stabilizethe system In this paper the LQR controller is used firstlyto control the front axle motor angle for the wheel synchro-nization and secondly to control the steering wheel motorfor a feedback torque
The wheel synchronization or directional control is abasic characteristic for steering function where a front tire
+
minus
sum
k1
k2
k3
x1
x2
x3
Plantr(t)
LQR controller
Figure 10 Basic control block diagram of LQR controller
angle should follow driver input steering wheel angle accord-ing to the steering ratio This is done by controlling theangle of the front axle motor In the LQR control schemesa feedback gain matrix is designed to achieve some compro-mise between the use of control effort the magnitude andthe speed of response that will guarantee a stable systemFigure 10 shows the control block diagram of the LQRcontroller
The optimal LQRproblem is often definedmore generallyand consists of finding controller transfer matrix that mini-mizes the cost function 119869 that is written as follows
119869 = int
prop
0(119909119879
119876119909+119906119879
119877119906) 119889119905 (31)
where 119909119879 and119880119879 respectively represent the transpose of the
state and input vector The 119876 and 119877 are positive definite (orpositive-semi definite) real symmetric matrix The 119876 and 119877
are usually chosen to be diagonal matrix written as follows
119876119894119894=
1maximum acceptable value of 1199092
119894
119877119895119895
=1
maximum acceptable value of 1199062119895
(32)
where
119894 = 120576 1 2 ℓ
119895 = 120576 1 2 119895 (33)
In essence Brysonrsquos rule scales the variables that appear in 119869 sothat the maximum acceptable value for each term is oneThisis especially important when the units used for the differentcomponents of 119906 and 119909 make the values for these variablesnumerically very different from each other
Assume that the steering ratio is 15 1 and the gains ofthe LQR controller are [01 098 0008] Figure 11 shows aresponse of the front axle system when the input steeringwheel angle is a step function as shown in Figure 11(a)
The results show that the angle of the front axle motorhas a fast response compared to the EPS steering system
International Journal of Vehicular Technology 9
0 2 4 6 8 100
2
4
6
8
10
12
14
16
18
20
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle
0 2 4 6 8 100
2
4
6
8
10
12
14
16
18
Time (s)
Mot
orc
olum
n an
gle (
deg)
EPS steeringmdashcolumn shaft angle
SBW systemmdashmotor angle
(b) Front axle motor angle
0 2 4 6 8 10
Rack
disp
lace
men
t (m
)
SBW system
EPS steering
Time (s)
18
16
14
12
10
8
6
4
2
0
times10minus3
(c) Rack displacement
0 2 4 6 8 100
02
04
06
08
1
12
14Fr
ont t
ire an
gle (
deg)
SBW system
EPS steering
Time (s)
(d) Front tire angle
Figure 11 Front axle system response for wheel synchronization
Moreover the LQR controller is able to reduce the delaycaused by the self aligning torque created from the contactbetween the tire and road surfaces This not only improvedthe vehicle stability but also ensured that the driver has amore confident level during manoeuvre Compared to theEPS steering system the delay caused by mechanical columnshaft connected to the steering wheel system Furthermore itcan be seen that the front tire angle is at almost fairly close to1∘ as shown in Figure 11(d) However it has small chatteringdue to the self aligning torque However with wide rangeadjustment gains of the LQR controller the steering responsecould be improved
32 Gain Scheduling (GS) with LQR Controller In certainsituations the dynamic behaviour of the process changed inaccordance with the conditions of the process This situation
could affect the stability of the system if not to control prop-erly Thus it is possible to change the gains of the controllerby monitoring the operating conditions of the process tohave a stable system The gain scheduling (GS) techniqueis used to change the gain of the LQR controller in order tocontrol the feedback torque and the steering feel gain (119896feel)is used to increase and decrease the torque It is important toincrease the feedback torque at high speed for steering stabil-ity and to decrease it in low speed especially during parkingto improve vehicle manoeuvre [9] This is based on the steer-ingwheel angle and vehicle speed function in order to achievethe feedback torque for the real driver steering feel
The relationship between a gain scheduling with LQRcontroller is shown in Table 4 The vehicle speed is dividedinto 3 categories at low speed (0ndash30 kmh) medium speed(31ndash100 kmh) and high speed (101ndash120 kmh) The maxi-mum vehicle speed is limited to 120 kmhThe steering wheel
10 International Journal of Vehicular Technology
0 2 4 6 8 10minus50
minus40
minus30
minus20
minus10
0
10
20
30
40
50
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Input steering wheel angle
0 1 2 3 4 5 6 7 8 9 10minus15
minus10
minus5
0
5
10
Time (s)
Stee
ring
torq
ue (N
m) LQR
LQR + GS
Reference
(b) 120583 = 1
0 1 2 3 4 5 6 7 8 9 10
minus15
minus10
minus5
0
5
10
15
Stee
ring
torq
ue (N
m) LQR
Reference
Time (s)
LQR + GS
(c) 120583 = 14
0 1 2 3 4 5 6 7 8 9 10minus20
minus15
minus10
minus5
0
5
10
15
20
Stee
ring
torq
ue (N
m) LQR LQR + GS
Reference
Time (s)
(d) 120583 = 17
Figure 12 Comparison of torque response of LQR and LQR + GS controller at different 120583
Table 4 The gains of LQR controller and steering feel
119881 (kmh) 120575119904119908
(degree) 1198961 1198962 1198963 119896feel
(0ndash30) (plusmn180∘ lt 120575119904119908
lt plusmn405∘) 003 065 0001 014(31ndash100) (plusmn46∘ lt 120575
119904119908lt plusmn179∘) 003 084 0001 060
(101ndash120) 120575119904119908
le +45∘ or 120575119904119908
le minus45∘ 005 095 0001 095
angle is also divided into 3 categories according to vehiclespeed condition The notation ldquo+rdquo means that the driverturned the steering wheel clockwise and notation ldquominusrdquo is usedfor counter clockwise in which the maximum steering wheelturn is plusmn405∘
The gains of the LQR controller are defined using Brysonrsquosrule method and the steering feel gain is defined usingintuitive and has a range from 01 to 1 This range has beendefined for various speeds 10 kmh 80 kmh and 120 kmhafter comparing with EPS steering system [15] For exampleif the driver turns a steering wheel between (plusmn180∘ lt 120575
119904119908lt
plusmn405∘) and vehicle speed is 10 kmh the total feedback torque
is multiplied by steering feel gain 119896feel = 014 Thus thefeedback torque is decreased which results in making thedriver feel it easier turning the steering wheel and improvesvehicle manoeuvre The feedback torque is regulated by LQRcontroller when the gains are 1198961 = 003 1198962 = 065 and1198963 = 0001 The gains are the optimum that has been assuredby providing a better torque control based on vehicle speedand steering wheel angle function
Comparing torque response between LQR and LQR +GScontroller for different coefficient (120583) is shown in Figure 12when the input steering wheel angle is a double lane changemanoeuvre as shown in Figure 12(a) The vehicle speed isset to be 80 kmh Based on results obtained the LQR + GStechnique offers an improvement 90 and the curve trend isalmost similar to reference torque compared to the LQR con-troller This is because the gains of LQR controller changedbased on the steering wheel angle and vehicle speed param-eter which is able to reject the uncertainty torque of a roadcondition Meanwhile with the fixed gains of the LQR con-troller it has an excessive torque that can cause the system tobe unstable The RMS torque values are illustrated in Table 5
International Journal of Vehicular Technology 11
Table 5 Comparison of RMS torque values of LQR and LQR + GScontroller
Reference LQR LQR + GS120583 = 1 56361 50795 55450120583 = 14 61218 68339 57243120583 = 17 66184 72888 62080
minus50 0 50minus15
minus10
minus5
0
5
10
15
Steering angle (deg)
Stee
ring
torq
ue (N
m) Feel gain = 5
Feel gain = 01Feel gain = 05
Feel gain = 1
(a) Different steering feel gain at 80 kmh
minus50 minus40 minus30 minus20 minus10 0 10 20 30 40 50
minus15
minus10
minus5
0
5
10
15
Steering wheel angle (deg)
Stee
ring
torq
ue (N
m)
80kmh kfeel = 060
120 kmh kfeel = 095
10kmh kfeel = 014
(b) Different steering feel gain at various speed
Figure 13 Steering feel gains response at various vehicle speed
Figure 13(a) shows the steering torque with respect tosteering wheel angle when different steering feel gain is usedat various vehicle speeds It is observed that when the steeringfeel gain is increased at maximum range which is 1 thesteering torque will increase and driver will sense stiffer onthe steering wheel at high speed manoeuvre Meanwhilewhen decreasing the steering feel gain this will result inreducing the steering torqueThus the driver will sense softer
Target monitor
Host PC
Cross over cableConnector block
PCI 6221
Power supplyCurrent sensor
DC motor driver
Figure 14 Steering wheel system hardware overview
during turning the steering wheel at the low speed manoeu-vre However when the steering feel gain is 5 which exceedsits maximum limit the steering torque becomes unstableFigure 13(b) shows the steering feel gains at various speedsAt high speed of 120 kmh the steering feel gain is increasedand this will increase the steering torque for steering stabilityMeanwhile at low speed the steering feel gain is decreasedwhich will help improve vehicle manoeuvre especially duringparking Based on the proposed gain scheduling with theLQR controller not only does it provide better control on thesteering torque but also it is optional for a designer to have awide range adjustment to vary the steering torque for a betterdriver steering feel and to improve vehicle manoeuvre
4 Hardware in the Loop (HIL) and Result
For hardware in the loop (HIL) setup the front axle systemand linear vehicle model are simulated on the host computerThe steering wheel system is replaced by HIL mechanismThe position of the steering wheel motor is measured usinga rotary encoder with 500 pulses per revolution To measurethe steering wheel torque the current sensor is connectedin series with the DC motor driver The steering wheel DCmotor has a torque 129Nm with 100 rpm The single tracklinear vehicle model is used to generate the estimation selfaligning torque as element to the feedback torque and distur-bance to the front axle system The HIL system is executedunder 1ms sampling time and interfaced using Matlab-XPCtarget toolbox configuration as shown in Figures 14 and 15
The assumptions through the experiment are listed asfollows
(i) the vehicle cruises at constant speed(ii) the condition of dry asphalt road (120583 = 085)
The experiments are done under high medium and lowspeed manoeuvre [26] The same input steering wheel anglebetween simulation and experiment is given Therefore theresult obtained from real time control is compared with theelectronic power steering (EPS) system [15]
41 Motor Model Validation for Steering Wheel SystemAlthough the model of steering wheel system has been
12 International Journal of Vehicular Technology
Stee
ring
whe
el sy
stem
DC motor
Current
Encoder
16 bit card PCI 6221
Analogoutput
Analoginput
Counterchannel
TargetPC
Monitortarget
PC
Crossover cable
Host PC
Front axle systemand
linear vehicle model
Hardware part Simulation part
Figure 15 Block diagram of the steering wheel-XPC target interfaced
Steering wheel model
120575m1
120575m1
ia1
r(t)
Figure 16 Block diagrammotormodel validation for steeringwheelsystem
derived the model validation is necessary to ensure thesystem is fairly able to adequate for real time applicationThis section describes the motor model validation of steeringwheel system The block diagram is shown in Figure 16
The input signal is 119903(119905) and the output is rate change ofmotor ( 120575
1198981) motor position (1205751198981) and the motor current
( 1198941198981) The measured output response is then used and com-
pared with simulation to estimate the parameter of the steer-ing wheel model using the Matlab tool software The mea-sured output data are taken based on the open loop systemwhen the input to the system is sinusoidal signal withamplitude of plusmn45∘ for 10 sec The Gaussian method is usedto estimate the parameters of the steering wheel motorThe comparison between simulation and measured outputis shown in Figure 17 It is obtained that the output of thesimulation is almost identical to the measured output How-ever small difference errors occur due to the small parametricuncertainties that have not been modelled The parameterestimation value is illustrated in Table 6
42 Noise Filtering Both analog and digital devices have atrait of noise or unwanted features that have effect on thesystem performance such as electromagnetic interferenceThe noise could be either random or white noise Attenuationof this noise is often as a primary goal in control systemapplication and in particular when controlling the processsystem The average value of noise typically is zero whichwill give misreading information to the process and it wouldnot be possible to control the process In order to reduce thenoise inference filtering technique is used A filter is used toremove unwanted noise It removes some noise frequenciesto suppress interfacing signal and reduce background noise
Table 6 Parameter estimation for DCmotor steering wheel system
Parameter Value Estimation value Unit1198771 4405 564 ohm1198711 0010 0017 henry119870119904119898
1459 0024 Nm1198691198981 0002 00036 (kgm2)1198871198981 0004 00068 Nm(rads)1198961199041 165 0025 Nm
There are many types of filtering techniques such as low passfilter that has been used in this study The low pass filter isused to filter out the interference noise from the output of thecurrent sensor
Figure 18 shows the comparison of the output responsefor a current sensor with and without using filters Withouta filter the current sensor has high amplitude interferencenoise However with low pass filter under the frequency of5Hz the interference noise is reduced and the fluctuations areremoved Thus this provides a smoother output response
43 Effect on Compensation Torque A steering torque com-parison without and with compensation torque is shown inFigure 19The results for without compensation torque showsthat the steering torque response curve is far away fromEPS steering system with different amplitude as shown inFigure 19(a) This happens due to less of derivative reactiontorque from the steering wheel motor Figure 19(b) showssteering torque response with compensation torque
It is observed that by adding the additional compensationtorque the steering torque response curve trends fairly closeto EPS steering system Thus it helps to improve the driversteering feel during manoeuvre and stabilize the systemHowever with the wide range adjustment gains of the com-pensation torque it could improve the steering torqueresponse
44 Medium Speed Manoeuvre The experimental steeringtorque results at medium speed manoeuvre are shown inFigure 20(b) when the steering wheel angle is plusmn100∘ as shown
International Journal of Vehicular Technology 13
0 2 4 6 8 10minus50
0
50
Time (s)
Rate
chan
ge m
otor
pos
ition
(deg
s)
MeasurementSimulation
(a) Rate change of motor position
0 2 4 6 8 100
10
20
30
40
50
60
Time (s)
Mot
or p
ositi
on (d
eg)
Measurement
Simulation
(b) Position of the motor
0 2 4 6 8 10minus01
minus005
0
005
01
015
Mot
or cu
rren
t (A
)
MeasurementSimulation
Time (s)
(c) Current of the motor
Figure 17 Motor model validation comparison between simulation and measurement
With filter Without filter
Time (s)
Curr
ent (
A)
008
006
004
002
0
minus002
minus004
minus006
minus0081 15 2 25 3 35 4 45 5 55
Figure 18 Noise filtering using low pass filter
in Figure 20(a) From the results the experimental steeringtorque response is almost identical to the EPS steering systemHowever it has small different torque amplitudes These
occur due to the fact that mechanism or friction from themechanical gearbox motor is not modelled in proposedcontrol algorithm
14 International Journal of Vehicular Technology
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
minus152 3 4 5 6 7 8 9 10
(a) Without compensation torque
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
21 3 4 5 6 7 8 9 10
(b) With compensation torque
Figure 19 Steering torque influence on compensation torque
minus80
minus60
minus40
minus20
0
20
40
60
80
100
Stee
ring
whe
el an
gle (
deg)
Time (s)3 4 5 6 7 8 9 10
(a) Steering wheel angle at 80 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
15
20
5
0
minus5
minus10
minus15
minus203 4 5 6 7 8 9 10
(b) Steering torque estimation at 80 kmh
Figure 20 Steering torque estimation and steering wheel angle response at 80 kmh
45 High Speed Manoeuvre At high speed manoeuvre theinput steering wheel angle that is more than plusmn45∘ is very dan-gerous [12] Therefore in this experiment the input steeringwheel angle is less than plusmn45∘ as shown in Figure 21(a) Basedon the results shown in Figures 21(b) and 21(c) the experi-mental steering torque is obtained fairly similar to the EPSsteering system Although the steering wheel angle is at smallangles the influence of vehicle speed cause the steering torqueis increased Moreover at higher speed the vehicle has moreunder steer propensity [9] This increased the self aligningtorque in proportional high torque at front axle motorThusthe driver will sense stiffer on the steering wheel duringmanoeuvre
In general the centering of the steering wheel perfor-mance or steering wheel returnability is better when thesteering torque gradient is large and the hysteresis is small [9]Based on this fact the result is shown in Figure 21(c) whereby
the hysteresis of the steering wheel returnability based onexperimental result is smaller compared to EPS steeringsystem In addition the steering torque gradient is larger thanEPS steering torque Therefore it is considered that the pro-posed feedback torque algorithmnot only achieved a steeringfeel but also improved the steering wheel returnability
46 Lower Speed Manoeuvre The steering torque results atlow speed manoeuvre are shown in Figures 22(b) and 22(c)when the input steering wheel is as shown in Figure 22(a)The vehicle cruised at constant speed of 10 kmh
The experimental steering torque is almost comparableto EPS torque but the amplitude is decreases The torque isdecreased due to the fact that the vehicle speed is reducedwhich is effect on reducing the demand to the total feedbacktorque Moreover the large turning of the steering wheelincreases the backlash of the motor which then contributes
International Journal of Vehicular Technology 15
2 3 4 5 6 7 8 9 10minus30
minus20
minus10
0
10
20
30
40
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle at 120 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steeringSimulationExperiment
10
5
0
minus5
minus10
3 4 5 6 7 8 9
(b) Steering torque estimation at 120 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment10
15
5
0
minus5
minus10
minus15minus20 minus10 0 10 20 30 40
Steering wheel angle (deg)
(c) Steering torque estimation against steering wheel angle at 120 kmh
Figure 21 Steering torque estimation response at 120 kmh
to the decrease of the motor torque Thus driver will sensesofter during steering the wheel during maneuver Howeverit has small different torque against EPS steering Wide rangeadjustment of gains in proposed control algorithm may helpimprove the steering torque response
3 4 5 6 7 8 9 10minus150
minus100
minus50
0
50
100
150
200
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
10
15
5
0
minus5
minus10
minus15
Time (s)2 3 4 5 6 7 8 9 10
(b) Steering torque estimation at 10 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
Steering wheel angle (deg)
10
8
6
4
2
0
minus2
minus4
minus6
minus8
minus10
minus100 minus50 0 50 100 150
(c) Steering torque estimation against steering wheel angle at 10 kmh
Figure 22 Steering torque estimation response at 10 kmh
5 Conclusion and Future Work
The proposed force feedback torque estimation and controlalgorithm have been described The proposed control algo-rithm is considered as an effort to generate the feedback
16 International Journal of Vehicular Technology
torque for driver steering feel in the vehicle SBW systemThe mathematical model of SBW subsystem consists ofsteering wheel and front axle systems that are modelledThe effectiveness of the control algorithm is verified throughthe experimental hardware in the loop (HIL) interfacedwith Matlab XPC target toolbox software According to theexperimental results the proposed control algorithm is ableto generate the feedback torque and the driver is able to sensesteering feel similar to that obtained in EPS steering systemThe LQR controller with gain scheduling based on steeringwheel angle and vehicle speed function provides better torquecontrol that allows rejection of the uncertainty torque froma road condition Furthermore the steering feel gain is ableto increase and lower the steering torque with improvedsteering stability at high speed and it improved the vehiclemanoeuvre at low speed It is also found that compensationtorque is able to improve the feedback torque and to stabilizethe system However it has small different torque magnitudethat occurs due to the backlash of steering wheel motor anduncertainty disturbance of mechanical properties Howeverwith a wide range adjustment of the gains on the proposedcontrol algorithm it may improve and provide realisticsteering torque for better driver steering feel
In future work the proposed control algorithm will betested in real vehicle in order to have more realistic steeringtorque for further investigation and analysis
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
Acknowledgments
This work is fully supported by the Ministry of Higher Edu-cation Malaysia and University Technology Malaysia underFRGS (vote 4F370) and Research University Grant (vote00G64) This work is also supported by Proton HoldingBerhad
References
[1] Y Yao ldquoVehicle steer-by-wire system controlrdquo SAE TechnicalPaper 2006-01-1175 2006
[2] E A Kumar D Dinesh and N Kamble ldquoAn overview of activefront steering systemrdquo International Journal of Scientific ampEngineering Research vol 3 no 6 2012
[3] S M H Fahami H Zamzuri S A Mazlan and N B Zulkar-nain ldquoThe design of vehicle active front steering based on steerby wire systemrdquo Advanced Science Letters vol 19 no 1 pp 61ndash65 2013
[4] S M H Fahami H Zamzuri S A Mazlan and S A SaruchildquoThe variable steering ratio for vehicle steer by wire systemusing hyperbolic tangent methodrdquo Applied Mechanics andMaterials vol 575 pp 781ndash784 2014
[5] T Kaufmann S Milsap B Murray and J Petrowski ldquoDevelop-ment experience with steer by wirerdquo in Proceedings of the SAEInternational Congress and Exhibition SAE 2001-01-2479 2001
[6] S Amberkare F Bolourchi J Demerly and S Millsap ldquoAcontrol system methodology for steer by wire systemsrdquo SAETechnical Paper 2004-01-1106 2004
[7] A Baviskar J R Wagner D M Dawson D Braganza andP Setlur ldquoAn adjustable steer-by-wire haptic-interface trackingcontroller for ground vehiclesrdquo IEEE Transactions on VehicularTechnology vol 58 no 2 pp 546ndash554 2009
[8] M Segawa S Kimura T Kada and S Nakona ldquoA study of reac-tive torque control for steer bywire systemrdquo inProceedings of theInternational Symposium on Advanced Vehicle Control (AVECrsquo02) pp 653ndash658 Hiroshima Japan September 2002
[9] S W Oh H C Chae S C Yun and C S Han ldquoThe design ofa controller for the steer-by-wire systemrdquo JSME InternationalJournal Series C Mechanical Systems Machine Elements andManufacturing vol 47 no 3 pp 896ndash907 2004
[10] C J Kim J H Jang S K Oh J Y Lee and J K HedrickldquoDevelopment of a control algorithm for a rack-actuating steer-by-wire system using road information feedbackrdquo Proceedingsof the Institution of Mechanical Engineers Part D Journal ofAutomobile Engineering vol 222 no 9 pp 1559ndash1571 2008
[11] S Asai H Kuroyanagi and S Takeuchi ldquoDevelopment of asteer-by-wire system with force feedback using a disturbanceobserverrdquo inProceedings of the SAEWorldCongressampExhibitionon Steering amp Suspension Technology Detroit Mich USA 2004
[12] T J Park C S Han and S H Lee ldquoDevelopment of the elec-tronic control unit for the rack-actuating steer-by-wire usingthe hardware-in-the-loop simulation systemrdquo Mechatronicsvol 15 no 8 pp 899ndash918 2005
[13] D Odenhtal T Bunte H D Heitzer and C Eicker ldquoHow tomake steer-by-wire feel like power steeringrdquo in Proceedings ofthe 15th IFACWorld Congress Barcelona Spain July 2000
[14] H Na C Zong and D Hu ldquoInvestigations on corneringcontrol algorithm design and road feeling optimization for asteer-by-wire vehiclerdquo in Proceedings of the IEEE InternationalConference on Mechatronics and Automation (ICMA rsquo09) pp3246ndash3251 Changchun China August 2009
[15] S Ancha A Baviskar J RWagner andDMDawson ldquoGroundvehicle steering systems modelling control and analysis ofhydraulic electric and steer-by-wire configurationsrdquo Interna-tional Journal of Vehicle Design vol 44 no 1-2 pp 188ndash2082007
[16] P Yih and J C Gerdes ldquoModification of vehicle handlingcharacteristics via steer-by-wirerdquo IEEE Transactions on ControlSystems Technology vol 13 no 6 pp 965ndash976 2005
[17] S M H Fahami H Zamzuri S A Mazlan and M A ZakarialdquoModeling and simulation of vehicle steer by wire systemrdquoin Proceedings of the IEEE Symposium on Humanities Scienceand Engineering Research (SHUSER rsquo12) pp 765ndash770 KualaLumpur Malaysia June 2012
[18] S Anwar and L Chen ldquoAn analytical redundancy-based faultdetection and isolation algorithm for a road-wheel controlsubsystem in a steer-by-wire systemrdquo IEEE Transactions onVehicular Technology vol 56 no 5 pp 2859ndash2869 2007
[19] M T Do Z Man C Zhang H Wang and F S Tay ldquoRobustsliding mode-based learning control for steer-by-wire systemsin modern vehiclesrdquo IEEE Transactions on Vehicular Technol-ogy vol 63 no 2 pp 580ndash590 2014
[20] J Duan R Wang and Y Yu ldquoResearch on control strategies ofsteer-by-wire systemrdquo in Proceedings of the International Con-ference on Intelligent Computation Technology and Automation(ICICTA rsquo10) vol 2 pp 1122ndash1125 May 2010
International Journal of Vehicular Technology 17
[21] B Zheng and S Anwar ldquoYaw stability control of a steer-by-wireequipped vehicle via active front wheel steeringrdquoMechatronicsvol 19 no 6 pp 799ndash804 2009
[22] S Sahoo S C Subramanian and S Srivastava ldquoDesign andimplementation of a controller for navigating an autonomousground vehiclerdquo in Proceedings of the 2nd International Confer-ence on Power Control and Embedded Systems (ICPCES rsquo12) pp1ndash6 IEEE Allahabad India December 2012
[23] T D Gillespie Fundamentals of Vehicle Dynamics Society ofAutomotive Engineers 1992
[24] M H Lee S K Ha J Y Choi and K S Yoon ldquoImprovement ofthe steering feel of an electric power steering system by torquemap modificationrdquo Journal of Mechanical Science and Technol-ogy vol 19 no 3 pp 792ndash801 2005
[25] R Pastorino M A Naya J A Perez and J Cuadrado ldquoGearedPM coreless motor modelling for driverrsquos force feedback insteer-by-wire systemsrdquo Mechatronics vol 21 no 6 pp 1043ndash1054 2011
[26] P Koehn andM Eckrich ldquoActive steeringmdashthe BMWapproachtowards modern steering technologyrdquo SAE Technical PaperSeries 2004-01-1105 SAE International 2004
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
2 International Journal of Vehicular Technology
Steering wheel system
Steering wheel
DC motor
Positionsensorencoder
ECU
Electronic controller unit Wiring
Front axle system
DC motorPositionsensorencoder
Rack and pinion Tire
Steer by wire (SBW) system
Steering wheel
CouplingMechanical column shaft
Conventional steering system
Figure 1 Difference between SBW system and conventional steering system
improvement to the torquemap has beenmade by adding thedamping torque to create more realistic driver steering feeland to improve steering wheel returnability [10] MeanwhileAmberkare et al [6] used the steering wheel angle and fed tomodel transfer function to create a feedback torque By usingthe appropriate transfer function it is possible to changethe desired steering feel by varying a proper gain of thetransfer function The approach of the disturbance observeris introduced by Asai et al [11] The position and current ofthe front axle motor are used to generate the feedback torqueFurthermore a reference model that consists of the inertiaand damping factor of steering wheel motor is proposed byPark et al [12] to generate the feedback torque The modelmatching technique which combines the mechanical andelectrical parts has been introduced by Odenhtal et al [13]to generate the feedback torque while the series of vehiclecornering control algorithm is taken into consideration tooptimize the feedback torque as proposed by Na et al [14]
The main objective of this paper is to proposes a controlalgorithm to generate the force feedback torque for driversteering feel in the vehicle SBW system The subsystem ofSBW system which is steering wheel and front axle system ismodelled and the control algorithm is explained To verify theeffectiveness of the proposed algorithm the hardware in theloop (HIL) is presented and interfaced using a Matlab XPCtarget toolbox The result was then is compared with electricpower steering (EPS) system [15]
2 The Modelling of Steer byWire (SBW) System
The SBW system is divided into two subsystems that consistof steering wheel and front axle system The system diagramof subsystem is shown in Figure 2
Table 1 Steering wheel system parameter
Parameter Description Value Unit1198771 Motor resistance 564 ohm1198711 Motor inductance 0017 henry119870119904119898
Motor constant 0024 Nm1198691198981 Motor inertia 00036 kgm2
1198871198981 Motor damping 00068 Nm(rads)1198961199041 Lumped torque stiffness 0025 Nm
1198811199041 Voltage source mdash Volt
21 Steering Wheel Model The main purpose of steeringwheel motor is to generate the steering feedback torque fordriver steering feel Figure 2(a) shows the system diagramfor the steering wheel system The input to the system is thetotal feedback torque (120591total feedback) and the output is the ratechange of steering wheel motor angular displacement ( 120575
1198981)the motor angular displacement (120575
1198981) and the current ofsteering wheel motor (119894
1198861) The parameters of the steeringwheel system are illustrated in Table 1
The mathematical equations of the steering wheel systemare written as follows
The steering motor angular displacement is
1205751198981 = minus
11988711989811198691198981
1205751198981 +
1198961199041
1198691198981
1198941198861 (1)
and the steering motor current is
1198941198861 = minus
11987711198711
1198941198861 +
119896119904119898
1198711
1205751198981 +
1198811199041
1198711 (2)
International Journal of Vehicular Technology 3
120591totalfeedback
ks1
120575sw
bm1
120575m1
Jm1
L1
R1
Vs1
ia1
(a)
Vs2
L2
R2
ia2
bm2
120575m2
Jm2
ks2
brackf
Mrackf
yrackfklf klfBkpf Bkpf
Jm2 Jm2
(b)
Figure 2 System diagram (a) steering wheel and (b) front axle system [15]
Consequently the state equation of steering wheel system isgiven as
119904119898
= 119860119904119898
119909119904119898
+119861119904119898
119906119904119898
119910119904119898
= 119862119904119898
119909119904119898
119909119904119898
= [ 1205751198981 1205751198981 1198941198861]
(3)
And 119906119904119898
is considered the input of the steering wheel motorand the parameter states are
119860119904119898
=
[[[[[[
[
minus(11988711989811198691198981
) 01198961199041
1198691198981
1 0 0
minus(119896119904119898
1198711) 0 minus(
11987711198711
)
]]]]]]
]
119861119904119898
=
[[[[
[
0
0
11198711
]]]]
]
119906119904119898
= 120591total feedback = 1198811199041
(4)
22 Front Wheel System The function of the front axle sys-tem is to ensure the front tire angle follows the steering wheelangle command according to the steering ratio [17]The frontaxle is composed of a front axle DC motor rack and pinionand tire model The system diagram of the front axle systemand the subsystem block diagram are shown in Figures 2(b)and 3 respectively The input to the system is the steeringwheel angle (120575
119904119908) while the front tire angle (120575
119891) is considered
the output The parameters of the front axle system areapproximated assumption that are illustrated in Table 2
Table 2 Front axle system parameter [15]
Parameter Description Value Unit1198772 Motor resistance 464 ohm1198712 Motor inductance 0015 henry119870119891119898
Motor constant 0032 Nm1198691198982 Motor inertia 00062 kgm2
1198871198982 Motor damping 00036 Nm(rads)1198961199042 Lumped torque stiffness 0025 Nm
119861rack Rack damping coefficient 0015 henry119872rack Rack lumped mass 0032 Nm119896119897119891
Rack linkage stiffness 000062 kgm2
119903119871
Offset of king pin axis 000036 m119903119901
Pinion gear radius 0025 m119861119896119901
King pin damping coefficient 000062 kgm2
119868119891
Lumped front wheel inertia 000036 kgm2
1198961199042 Lumped torque stiffness 0025 Nm
120575sw 120575m2 120575fyrack
Front axle system
Front axleDC motor
Rack andpinionsystem
Front tiredynamics
Figure 3 Block diagram of front axle subsystem
The mathematical equation of the front axle DC motor iswritten as follows
The front axle motor angular displacement is
1205751198982 = minus
11988711989821198691198982
1205751198982 +
1198961199042
1198691198982
1198941198862 (5)
4 International Journal of Vehicular Technology
and the front axle motor current is
1198941198862 = minus
11987721198712
1198941198862 +
119896119891119898
1198712
1205751198982 +
1198811199042
1198712 (6)
Consequently the state equations of the front axle motor aregiven as
119891119898
= 119860119891119898
119909119891119898
+119861119891119898
119906119891119898
119910119891119898
= 119862119891119898
119909119891119898
119909119891119898
= [ 1205751198982 1205751198982 1198941198862]
(7)
And 119906119891119898
is the input to the front axle motor while outputis the rate change of front axle motor angular displacement( 1205751198982) the front axle motor angular displacement (120575
1198982) andthe current of front axle motor (119894
1198862)The parameter states are written as follows
119860119891119898
=
[[[[[[[
[
minus(11988711989821198691198982
) 01198961199042
1198691198982
1 0 0
minus(
119896119891119898
1198712) 0 minus(
11987721198712
)
]]]]]]]
]
119861119891119898
=
[[[[
[
0
0
11198712
]]]]
]
119906119891119898
= 120575119904119908
= 1198811199042
(8)
The rack and pinion system is
119910rack =1
119872rack[(minus
2119896119897119891119910rack
1199032119871
)minus(1198961199042119910rack1199032119901
)
minus119861rack119891 119910rack +(11989611990421205751198982119903119901
)]
(9)
And 119906119910rack is considered the input of the rack and pinion
system and the output is the rack displacement (119910rack) andthe parameter states are as follows
119860119910rack
=[[[
[
0 1
minus[(
2119896119897119891
119872rack1199032119871
) + (1198961199042
119872rack1199032119901
)] minus(119861rack119872rack
)
]]]
]
119861119910rack =
[[
[
0
(1198961199042
119872rack119903119901)
]]
]
119906119910rack = 120575
1198982
(10)
Fyr
lr lf
r
120572r
120573
Fyf120572f
120575f
Figure 4 Single track-linear vehicle model
and the front tire dynamic is
120575119891=
1119868119891
[(minus119896119897119891120575119891+(
119896119897119891119910rack
119903119871
)minus119861119896119901
120575119891)] (11)
and 119906tire is considered the input of the front tire dynamicTheoutput is the front tire angle (120575
119891) and the parameter state is
119860 tire =[[
[
0 1
minus(
119896119897119891
119868119891
) minus(
119861119896119901
119868119891
)
]]
]
119861tire =[[
[
0
(
119896119897119891
119868119891119903119871
)
]]
]
119906tire = 119910rack
(12)
23 Single Track-Linear Vehicle Model The stability charac-teristics of a steer-by-wire vehicle are affected by the influenceof front tires via the feedback of the vehicle dynamic responseThus it is necessary to develop a single track-linear vehiclemodel as mentioned by author [16 18ndash21] in simulation andexperimental validationThe single track linear vehiclemodelis shown in Figure 4 In this paper the aim is to model thelinear vehicle in order to generate the estimation of self align-ing torque (120591
119886) This acts as the disturbance input to the front
axle system and assures the road contact with front tire Thedynamics of single track-linear vehicle model is representedby bicycle model where the states contain yaw rate (119903) andvehicle body slip angle (120573)
The following assumptions are considered for normaldriving maneuverer These assumptions could simplify thevehicle model [18]
(i) friction force in the 119909-direction is negligible
(ii) vehicle speed is constant
(iii) angle for right and left tire is approximately the same
International Journal of Vehicular Technology 5
With the aforementioned assumptions the derived equationsof the motion of yaw rate (119903) and vehicle body slip angle (120573)are shown in (13) as follows
119903 = [
[
(1198972119891119862119891+ 119897
2119903119862119903)
119868119904119881
]
]
119903+[
[
(1198972119903119862119903minus 119897
2119891119862119891)
119868119904
]
]
120573
+[
(119897119891119862119891)
119868119904
]120575119891
120573 = [1+(119897119903119862119903minus 119897119891119862119891)
1198981198812 ] 119903 minus[
(119862119891+ 119862119903)
119898119881]120573
+[
119862119891
119898119881]120575119891
(13)
In the linear tire region lateral force at the front and rear isrelated to slip angle by cornering stiffness of the front and reartires written as follows
119865119910119891
= minus119862119891120572119891
119865119910119903
= minus119862119903120572119903
(14)
At a larger slip angle the relation between lateral force andslip angle becomes nonlinear By assuming a smaller slipangle approximation the slip angle is written in terms ofvehicle body slip angle and yaw rate that are considefed asfollows
120572119891= (120573+
119897119891
119881119903minus 120575119891)
120572119903= (120573+
119897119903
119881119903minus 120575119903)
(15)
Rewriting (13) into state space form we haveV119898 = 119860V119898119909V119898 +119861V119898119906V119898
119910V119898 = 119862V119898119909V119898
119909V119898 = [119903 120573]
(16)
And 119906V119898 is considered the input to the linear vehicle modelwhile the yaw rate (119903) and vehicle body slip angle (120573) arethe output of the linear vehicle model The parameter stateis written as follows119860V119898
=
[[[[[[
[
(
(1198712119891119862119891+ 119871
2119903119862119903)
119868119904119881
) minus
(119871119903119862119903minus 119871119891119862119891)
119868119904
1 + (
(119871119903119862119903minus 119871119891119862119891)
1198981198812 ) minus
(119862119891+ 119862119903)
119898119881
]]]]]]
]
119861V119898 =[[[
[
119871119891119862119891
119868119904
119862119891
119898119881
]]]
]
119906V119898 = 120575119891
(17)
The parameter of a linear vehicle model directly can bemeasured but some parameters need to be estimated Thereare several methods to estimate themodel parameter [22 23]A method based on Sahoo et al [22] is used to estimate themodel parameters with assumption that the vehicle mass isdistributed to front and rear of each tire The total of vehiclemass (119898) that is based on the front (119898
119891) and rear (119898
119903) tire
masses is written as follows
119898 = 119898119891+119898119903
(18)
when
119898119891= 119898119891119897+119898119891119903
119898119903= 119898119903119897+119898119903119903
(19)
where 119898119891119897 119898119891119903 119898119903119897 and 119898
119903119903are the masses of the vehicle
at front left front right rear left and rear right respectivelyThe parameter length of wheel base (119897) which is sum of lengthfrom a center of front tire 119897
119891and rear tire 119897
119903to COG of vehicle
body is written as follows
119897 = 119897119891+ 119897119903 (20)
while the inertia of vehicle 119868119904that is approximated by assum-
ing the vehicle as two point masses joined with mass-lessroad [22] is written as follows
119868119904= 1198981198911198972119891+1198981199031198972119903 (21)
The front cornering stiffness 119862119891and rear cornering stiffness
119862119903parameters are written in (22) by assuming that the type
of radial tire is used and cornering stiffness per degree of slipangle is approximately 16-17 of the load on the tire [22]Consider
119862119891= 119898119891lowast 119892 lowast 0165 (Ndeg)
119862119903= 119898119903lowast 119892 lowast 0165 (Ndeg)
(22)
The parameters of the linear vehicle model are illustrated inTable 3
3 Force Feedback Torque EstimationControl Algorithm
It has been known that the force feedback torque for driversteering feel should be created artificially due to the elimina-tion of mechanical column shaft in SBW system
Moreover the proportion of the steering feel obtainedshould be similar to conventional steering system For thisreason the steering wheel in SBW system that is equippedwith the motor actuator is used to generate the steeringfeel by controlling the total feedback torque (120591total feedback)A schematic block diagram of the proposed force feedbacktorque control is shown in Figure 5
Based on Figure 5 the total force feedback torque consistsof the torque of front axle motor (120591
119891119898) torque on steering
wheel motor (120591119904119898) and compensation torque which is the
inertia (120591inertia) and damping torque (120591damp) Meanwhile the
6 International Journal of Vehicular Technology
+
+
+
minus
minus
Wheel synchronization
120575sw
120575sw
120591a
Frontaxle
system
LQRcontroller
LQRcontroller
120575m2
120575m2
120575f
ia2
120573
r
Linearvehiclemodel
Force feedback torque
120591map
120591fm
120591sm
120591damp
120591inertia
VGain
scheduling
kfm
kfeelsum
120591total feedback Steeringwheelsystem
Steering wheel system
Steeringwheel
120575m1
120575m1
ia1
ksm
Estimation selfaligning torque
Torque(RMS)
Figure 5 Overview of force feedback torque control algorithm for SBW system
Table 3 Linear vehicle model [16]
Parameter Description Value Unit119862119891 Front cornering stiffness 40000 Nrad
119862119903 Rear cornering stiffness 35000 Nrad
119897119891
Length from center frontwheel to (COG) vehicle 14 m
119897119903
Length from center rearwheel to COG vehicle 1 m
119898 Mass of the vehicle 1535 kg119868119904 Inertia of vehicle 2149 kgm2
119881 Speed of vehicle mdash kmh119865119910119891 Front tire lateral force mdash Nm
119865119910119903 Rear tire lateral force mdash Nm
120572119891 Front tire slip angle mdash Nm
120572119903 Rear tire slip angle mdash Nm
torque map is based on the function of the steering wheelangle and vehicle speed is added to the total feedback torqueMoreover a steering feel gain (119896feel) is adapted based on thefunction of steering wheel angle and vehicle speed to varythe feedback torque In order to control the feedback torquethe gain scheduling with LQR controller is used To showthe influence of the tire on the road surface the self aligningtorque (120591
119886) acts as input disturbance to the front axle system
as follows
120591119886= minus119862
119891(119905119901+ 119905119898) [120573+
119897119891119903
119881minus 120575119891]120583 (23)
where 120583 is coefficient for road dry condition pneumatic trail(119905119901) and mechanical trail (119905
119898)
On the other hand any change of the road surface couldchange the feedback torque magnitude of the front axlemotor Furthermore the LQR controller of the front axle sys-tem is used to synchronize the front tire angle (120575
119891) with driver
input steering wheel angle (120575119904119908) by controlling the angle of
the front axle motor (1205751198982) in accordance with the steering
ratioThus based on the proposed control algorithm the total
force feedback torque to the steering wheel system is writtenas follows
120591total feedback = ([(radicmean (120591119904119898
+ 120591119891119898
+ 120591map))]
+ 120591inertia + 120591damp) 119896feel
(24)
Author [9 10] has generated and discussed the force feed-back torque using the torque map conceptThe torque map isbased on the function of the steering wheel angle and vehiclespeed parameter The feedback torque is determined by twocontrol parameters (119896
120573) for the vehicle speed and (119896
120572) for the
steering wheel angle functionThe torque based on the function of vehicle speed is given
as follows [9]
1205911 = 1198961205731199092(13119909minus
12119881max)+119879in (25)
International Journal of Vehicular Technology 7
0 2 4 6 8 10minus40
minus30
minus20
minus10
0
10
20
30
40
Time (s)
Stee
ring
torq
ue (N
m)
70kmh75kmh
50kmh40kmh30kmh20kmh10kmh
80kmh
60kmh
Figure 6 Steering torque based on torque map at different speed
where (119909) is initial velocity and (119881max) is maximum velocitywhile the torque based on a function of steering wheel angleis written as follows [9]
1205912 = 119896120572radic1003816100381610038161003816120575119904119908
1003816100381610038161003816 sgn 120575119904119908 (26)
Based on the torque map concept the results of steeringtorque at different vehicle speed are shown in Figure 6 whenthe input steering wheel angle is a double lane changemanoeuvreThe torque increased at high speed anddecreasedat low speed [9 24] By varying the two control parametersthe steering reactive torque changes at different torque mag-nitude However the feedback torque based on the torquemap concept did not create the artificial steering feel dueto the fact that no element of road contact such as vehicledynamic responses are feedback to the torque map modelThus driver not sense realistic driver steering feel
Thus for this reason the front axle motor torque andsteering wheel motor torque are added to the torque mapThese ensure that the driver senses the feedback torque fromroad contact The front axle motor torque is chosen due toits proportion to the road surface The reaction torque of theroad surface gives the effect to the rack and pinion systemand it connects to the front axle DC motor Thus the frontaxle motor torque would always be proportional to the roadsurface Any changes in road surface will influence the frontaxle motor torque In order to measure the torque at the frontaxle motor the direct currentmeasurement technique is usedto estimate the torque
The direct relationship between torque and current of amotor can be defined [25]Thus then front axlemotor torque(120591119891119898
) that consists of the front axle motor constant (119896119891119898
) andthe motor current (119894
1198862) is written as follows
120591119891119898
= 1198941198862119896119891119898 (27)
and the corresponding steeringwheelmotor torque (120591119904119898) that
consists of the steering wheel motor constant (119896119904119898) and the
motor current (1198941198861) is written as follows
120591119904119898
= 1198941198861119896119904119898 (28)
0 2 4 6 8 10Time (s)
Mot
or to
rque
(Nm
)
Front axle motor torque
Steering wheel motor torque6
4
2
0
minus2
minus4
minus6
Figure 7 Steering wheel and front axle motor torque at 80 kmh
Figure 7 shows the front axle motor and steering wheelmotor torque response when the input is a double lanechangemanoeuvreThe results show that the front axlemotorgives a high torque This is due to the fact that the inputdisturbance torque between the tire and road surface which isthe self aligning torque has an effect on the front axle motorIncreasing the vehicle speedwill increase the front axlemotortorque On the other hand the magnitude of steering wheelmotor torque is decreased due to the elimination of mechan-ical column shaft and is not affected by the self aligningtorque
The phase compensation torque consists of the inertiatorque and damping torque potential to recreate the artificialsteering feel of the conventional steering system Moreoverit is intended to adjust the steering feel reduce the vibrationand stabilize the system Due to this fact the element of com-pensation torque is taken in order to improve the feedbacktorque for a driver steering feel and to stabilize the system
The damping torque (120591damp) which is the rate change ofthe steering wheel motor angle ( 120575
1198981) and the damping gain(119896damp) is written as follows
120591damp = 1205751198981119896damp (29)
The inertia torque (120591inertia) that is the acceleration of thesteering wheel motor angle ( 120575
1198981) and the inertia gain (119896inertia)is written as follows
120591inertia = 1205751198981119896inertia (30)
The inertia and damping torque responses are shown inFigure 8 when the input steering wheel angle is a double lanechange manoeuvre It is obtained that the damping torquegives a smooth signal and reduces the vibration Meanwhilethe inertia torque gives high magnitude Thus the combina-tion between the inertia torque and damping torque not onlyprovides an average of steering feel but is also able to stabilizethe system at the same time Figure 9 shows the feedback
8 International Journal of Vehicular Technology
0 2 4 6 8 10minus6
minus4
minus2
0
2
4
6
Time (s)
Torq
ue (N
m)
Inertia torque Damping torque
Figure 8 Inertia and damping torque
0 2 4 6 8 10Time (s)
Forc
e fee
dbac
k to
rque
(Nm
)
15
10
5
0
minus5
minus10
minus15
minus20
minus25
10kmh
50kmh100 kmh120 kmh
135kmh
Figure 9 Proposed force feedback torque response at differentvehicle speed
torque response based on (24) The torque is increasedaccording to the vehicle speed The higher vehicle speed willincrease the feedback torque and vice versa due to the samefact that is proposed by torque map
31 Linear Quadratic Regulator (LQR) Controller The LQRcontroller provides the best possible performance to controland stabilize the system by changing the location of the polesof the system to the optimal location for a time responseovershoot and steady stateThis is done by designing the statefeedback control119870 thatminimized cost function 119869 to stabilizethe system In this paper the LQR controller is used firstlyto control the front axle motor angle for the wheel synchro-nization and secondly to control the steering wheel motorfor a feedback torque
The wheel synchronization or directional control is abasic characteristic for steering function where a front tire
+
minus
sum
k1
k2
k3
x1
x2
x3
Plantr(t)
LQR controller
Figure 10 Basic control block diagram of LQR controller
angle should follow driver input steering wheel angle accord-ing to the steering ratio This is done by controlling theangle of the front axle motor In the LQR control schemesa feedback gain matrix is designed to achieve some compro-mise between the use of control effort the magnitude andthe speed of response that will guarantee a stable systemFigure 10 shows the control block diagram of the LQRcontroller
The optimal LQRproblem is often definedmore generallyand consists of finding controller transfer matrix that mini-mizes the cost function 119869 that is written as follows
119869 = int
prop
0(119909119879
119876119909+119906119879
119877119906) 119889119905 (31)
where 119909119879 and119880119879 respectively represent the transpose of the
state and input vector The 119876 and 119877 are positive definite (orpositive-semi definite) real symmetric matrix The 119876 and 119877
are usually chosen to be diagonal matrix written as follows
119876119894119894=
1maximum acceptable value of 1199092
119894
119877119895119895
=1
maximum acceptable value of 1199062119895
(32)
where
119894 = 120576 1 2 ℓ
119895 = 120576 1 2 119895 (33)
In essence Brysonrsquos rule scales the variables that appear in 119869 sothat the maximum acceptable value for each term is oneThisis especially important when the units used for the differentcomponents of 119906 and 119909 make the values for these variablesnumerically very different from each other
Assume that the steering ratio is 15 1 and the gains ofthe LQR controller are [01 098 0008] Figure 11 shows aresponse of the front axle system when the input steeringwheel angle is a step function as shown in Figure 11(a)
The results show that the angle of the front axle motorhas a fast response compared to the EPS steering system
International Journal of Vehicular Technology 9
0 2 4 6 8 100
2
4
6
8
10
12
14
16
18
20
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle
0 2 4 6 8 100
2
4
6
8
10
12
14
16
18
Time (s)
Mot
orc
olum
n an
gle (
deg)
EPS steeringmdashcolumn shaft angle
SBW systemmdashmotor angle
(b) Front axle motor angle
0 2 4 6 8 10
Rack
disp
lace
men
t (m
)
SBW system
EPS steering
Time (s)
18
16
14
12
10
8
6
4
2
0
times10minus3
(c) Rack displacement
0 2 4 6 8 100
02
04
06
08
1
12
14Fr
ont t
ire an
gle (
deg)
SBW system
EPS steering
Time (s)
(d) Front tire angle
Figure 11 Front axle system response for wheel synchronization
Moreover the LQR controller is able to reduce the delaycaused by the self aligning torque created from the contactbetween the tire and road surfaces This not only improvedthe vehicle stability but also ensured that the driver has amore confident level during manoeuvre Compared to theEPS steering system the delay caused by mechanical columnshaft connected to the steering wheel system Furthermore itcan be seen that the front tire angle is at almost fairly close to1∘ as shown in Figure 11(d) However it has small chatteringdue to the self aligning torque However with wide rangeadjustment gains of the LQR controller the steering responsecould be improved
32 Gain Scheduling (GS) with LQR Controller In certainsituations the dynamic behaviour of the process changed inaccordance with the conditions of the process This situation
could affect the stability of the system if not to control prop-erly Thus it is possible to change the gains of the controllerby monitoring the operating conditions of the process tohave a stable system The gain scheduling (GS) techniqueis used to change the gain of the LQR controller in order tocontrol the feedback torque and the steering feel gain (119896feel)is used to increase and decrease the torque It is important toincrease the feedback torque at high speed for steering stabil-ity and to decrease it in low speed especially during parkingto improve vehicle manoeuvre [9] This is based on the steer-ingwheel angle and vehicle speed function in order to achievethe feedback torque for the real driver steering feel
The relationship between a gain scheduling with LQRcontroller is shown in Table 4 The vehicle speed is dividedinto 3 categories at low speed (0ndash30 kmh) medium speed(31ndash100 kmh) and high speed (101ndash120 kmh) The maxi-mum vehicle speed is limited to 120 kmhThe steering wheel
10 International Journal of Vehicular Technology
0 2 4 6 8 10minus50
minus40
minus30
minus20
minus10
0
10
20
30
40
50
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Input steering wheel angle
0 1 2 3 4 5 6 7 8 9 10minus15
minus10
minus5
0
5
10
Time (s)
Stee
ring
torq
ue (N
m) LQR
LQR + GS
Reference
(b) 120583 = 1
0 1 2 3 4 5 6 7 8 9 10
minus15
minus10
minus5
0
5
10
15
Stee
ring
torq
ue (N
m) LQR
Reference
Time (s)
LQR + GS
(c) 120583 = 14
0 1 2 3 4 5 6 7 8 9 10minus20
minus15
minus10
minus5
0
5
10
15
20
Stee
ring
torq
ue (N
m) LQR LQR + GS
Reference
Time (s)
(d) 120583 = 17
Figure 12 Comparison of torque response of LQR and LQR + GS controller at different 120583
Table 4 The gains of LQR controller and steering feel
119881 (kmh) 120575119904119908
(degree) 1198961 1198962 1198963 119896feel
(0ndash30) (plusmn180∘ lt 120575119904119908
lt plusmn405∘) 003 065 0001 014(31ndash100) (plusmn46∘ lt 120575
119904119908lt plusmn179∘) 003 084 0001 060
(101ndash120) 120575119904119908
le +45∘ or 120575119904119908
le minus45∘ 005 095 0001 095
angle is also divided into 3 categories according to vehiclespeed condition The notation ldquo+rdquo means that the driverturned the steering wheel clockwise and notation ldquominusrdquo is usedfor counter clockwise in which the maximum steering wheelturn is plusmn405∘
The gains of the LQR controller are defined using Brysonrsquosrule method and the steering feel gain is defined usingintuitive and has a range from 01 to 1 This range has beendefined for various speeds 10 kmh 80 kmh and 120 kmhafter comparing with EPS steering system [15] For exampleif the driver turns a steering wheel between (plusmn180∘ lt 120575
119904119908lt
plusmn405∘) and vehicle speed is 10 kmh the total feedback torque
is multiplied by steering feel gain 119896feel = 014 Thus thefeedback torque is decreased which results in making thedriver feel it easier turning the steering wheel and improvesvehicle manoeuvre The feedback torque is regulated by LQRcontroller when the gains are 1198961 = 003 1198962 = 065 and1198963 = 0001 The gains are the optimum that has been assuredby providing a better torque control based on vehicle speedand steering wheel angle function
Comparing torque response between LQR and LQR +GScontroller for different coefficient (120583) is shown in Figure 12when the input steering wheel angle is a double lane changemanoeuvre as shown in Figure 12(a) The vehicle speed isset to be 80 kmh Based on results obtained the LQR + GStechnique offers an improvement 90 and the curve trend isalmost similar to reference torque compared to the LQR con-troller This is because the gains of LQR controller changedbased on the steering wheel angle and vehicle speed param-eter which is able to reject the uncertainty torque of a roadcondition Meanwhile with the fixed gains of the LQR con-troller it has an excessive torque that can cause the system tobe unstable The RMS torque values are illustrated in Table 5
International Journal of Vehicular Technology 11
Table 5 Comparison of RMS torque values of LQR and LQR + GScontroller
Reference LQR LQR + GS120583 = 1 56361 50795 55450120583 = 14 61218 68339 57243120583 = 17 66184 72888 62080
minus50 0 50minus15
minus10
minus5
0
5
10
15
Steering angle (deg)
Stee
ring
torq
ue (N
m) Feel gain = 5
Feel gain = 01Feel gain = 05
Feel gain = 1
(a) Different steering feel gain at 80 kmh
minus50 minus40 minus30 minus20 minus10 0 10 20 30 40 50
minus15
minus10
minus5
0
5
10
15
Steering wheel angle (deg)
Stee
ring
torq
ue (N
m)
80kmh kfeel = 060
120 kmh kfeel = 095
10kmh kfeel = 014
(b) Different steering feel gain at various speed
Figure 13 Steering feel gains response at various vehicle speed
Figure 13(a) shows the steering torque with respect tosteering wheel angle when different steering feel gain is usedat various vehicle speeds It is observed that when the steeringfeel gain is increased at maximum range which is 1 thesteering torque will increase and driver will sense stiffer onthe steering wheel at high speed manoeuvre Meanwhilewhen decreasing the steering feel gain this will result inreducing the steering torqueThus the driver will sense softer
Target monitor
Host PC
Cross over cableConnector block
PCI 6221
Power supplyCurrent sensor
DC motor driver
Figure 14 Steering wheel system hardware overview
during turning the steering wheel at the low speed manoeu-vre However when the steering feel gain is 5 which exceedsits maximum limit the steering torque becomes unstableFigure 13(b) shows the steering feel gains at various speedsAt high speed of 120 kmh the steering feel gain is increasedand this will increase the steering torque for steering stabilityMeanwhile at low speed the steering feel gain is decreasedwhich will help improve vehicle manoeuvre especially duringparking Based on the proposed gain scheduling with theLQR controller not only does it provide better control on thesteering torque but also it is optional for a designer to have awide range adjustment to vary the steering torque for a betterdriver steering feel and to improve vehicle manoeuvre
4 Hardware in the Loop (HIL) and Result
For hardware in the loop (HIL) setup the front axle systemand linear vehicle model are simulated on the host computerThe steering wheel system is replaced by HIL mechanismThe position of the steering wheel motor is measured usinga rotary encoder with 500 pulses per revolution To measurethe steering wheel torque the current sensor is connectedin series with the DC motor driver The steering wheel DCmotor has a torque 129Nm with 100 rpm The single tracklinear vehicle model is used to generate the estimation selfaligning torque as element to the feedback torque and distur-bance to the front axle system The HIL system is executedunder 1ms sampling time and interfaced using Matlab-XPCtarget toolbox configuration as shown in Figures 14 and 15
The assumptions through the experiment are listed asfollows
(i) the vehicle cruises at constant speed(ii) the condition of dry asphalt road (120583 = 085)
The experiments are done under high medium and lowspeed manoeuvre [26] The same input steering wheel anglebetween simulation and experiment is given Therefore theresult obtained from real time control is compared with theelectronic power steering (EPS) system [15]
41 Motor Model Validation for Steering Wheel SystemAlthough the model of steering wheel system has been
12 International Journal of Vehicular Technology
Stee
ring
whe
el sy
stem
DC motor
Current
Encoder
16 bit card PCI 6221
Analogoutput
Analoginput
Counterchannel
TargetPC
Monitortarget
PC
Crossover cable
Host PC
Front axle systemand
linear vehicle model
Hardware part Simulation part
Figure 15 Block diagram of the steering wheel-XPC target interfaced
Steering wheel model
120575m1
120575m1
ia1
r(t)
Figure 16 Block diagrammotormodel validation for steeringwheelsystem
derived the model validation is necessary to ensure thesystem is fairly able to adequate for real time applicationThis section describes the motor model validation of steeringwheel system The block diagram is shown in Figure 16
The input signal is 119903(119905) and the output is rate change ofmotor ( 120575
1198981) motor position (1205751198981) and the motor current
( 1198941198981) The measured output response is then used and com-
pared with simulation to estimate the parameter of the steer-ing wheel model using the Matlab tool software The mea-sured output data are taken based on the open loop systemwhen the input to the system is sinusoidal signal withamplitude of plusmn45∘ for 10 sec The Gaussian method is usedto estimate the parameters of the steering wheel motorThe comparison between simulation and measured outputis shown in Figure 17 It is obtained that the output of thesimulation is almost identical to the measured output How-ever small difference errors occur due to the small parametricuncertainties that have not been modelled The parameterestimation value is illustrated in Table 6
42 Noise Filtering Both analog and digital devices have atrait of noise or unwanted features that have effect on thesystem performance such as electromagnetic interferenceThe noise could be either random or white noise Attenuationof this noise is often as a primary goal in control systemapplication and in particular when controlling the processsystem The average value of noise typically is zero whichwill give misreading information to the process and it wouldnot be possible to control the process In order to reduce thenoise inference filtering technique is used A filter is used toremove unwanted noise It removes some noise frequenciesto suppress interfacing signal and reduce background noise
Table 6 Parameter estimation for DCmotor steering wheel system
Parameter Value Estimation value Unit1198771 4405 564 ohm1198711 0010 0017 henry119870119904119898
1459 0024 Nm1198691198981 0002 00036 (kgm2)1198871198981 0004 00068 Nm(rads)1198961199041 165 0025 Nm
There are many types of filtering techniques such as low passfilter that has been used in this study The low pass filter isused to filter out the interference noise from the output of thecurrent sensor
Figure 18 shows the comparison of the output responsefor a current sensor with and without using filters Withouta filter the current sensor has high amplitude interferencenoise However with low pass filter under the frequency of5Hz the interference noise is reduced and the fluctuations areremoved Thus this provides a smoother output response
43 Effect on Compensation Torque A steering torque com-parison without and with compensation torque is shown inFigure 19The results for without compensation torque showsthat the steering torque response curve is far away fromEPS steering system with different amplitude as shown inFigure 19(a) This happens due to less of derivative reactiontorque from the steering wheel motor Figure 19(b) showssteering torque response with compensation torque
It is observed that by adding the additional compensationtorque the steering torque response curve trends fairly closeto EPS steering system Thus it helps to improve the driversteering feel during manoeuvre and stabilize the systemHowever with the wide range adjustment gains of the com-pensation torque it could improve the steering torqueresponse
44 Medium Speed Manoeuvre The experimental steeringtorque results at medium speed manoeuvre are shown inFigure 20(b) when the steering wheel angle is plusmn100∘ as shown
International Journal of Vehicular Technology 13
0 2 4 6 8 10minus50
0
50
Time (s)
Rate
chan
ge m
otor
pos
ition
(deg
s)
MeasurementSimulation
(a) Rate change of motor position
0 2 4 6 8 100
10
20
30
40
50
60
Time (s)
Mot
or p
ositi
on (d
eg)
Measurement
Simulation
(b) Position of the motor
0 2 4 6 8 10minus01
minus005
0
005
01
015
Mot
or cu
rren
t (A
)
MeasurementSimulation
Time (s)
(c) Current of the motor
Figure 17 Motor model validation comparison between simulation and measurement
With filter Without filter
Time (s)
Curr
ent (
A)
008
006
004
002
0
minus002
minus004
minus006
minus0081 15 2 25 3 35 4 45 5 55
Figure 18 Noise filtering using low pass filter
in Figure 20(a) From the results the experimental steeringtorque response is almost identical to the EPS steering systemHowever it has small different torque amplitudes These
occur due to the fact that mechanism or friction from themechanical gearbox motor is not modelled in proposedcontrol algorithm
14 International Journal of Vehicular Technology
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
minus152 3 4 5 6 7 8 9 10
(a) Without compensation torque
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
21 3 4 5 6 7 8 9 10
(b) With compensation torque
Figure 19 Steering torque influence on compensation torque
minus80
minus60
minus40
minus20
0
20
40
60
80
100
Stee
ring
whe
el an
gle (
deg)
Time (s)3 4 5 6 7 8 9 10
(a) Steering wheel angle at 80 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
15
20
5
0
minus5
minus10
minus15
minus203 4 5 6 7 8 9 10
(b) Steering torque estimation at 80 kmh
Figure 20 Steering torque estimation and steering wheel angle response at 80 kmh
45 High Speed Manoeuvre At high speed manoeuvre theinput steering wheel angle that is more than plusmn45∘ is very dan-gerous [12] Therefore in this experiment the input steeringwheel angle is less than plusmn45∘ as shown in Figure 21(a) Basedon the results shown in Figures 21(b) and 21(c) the experi-mental steering torque is obtained fairly similar to the EPSsteering system Although the steering wheel angle is at smallangles the influence of vehicle speed cause the steering torqueis increased Moreover at higher speed the vehicle has moreunder steer propensity [9] This increased the self aligningtorque in proportional high torque at front axle motorThusthe driver will sense stiffer on the steering wheel duringmanoeuvre
In general the centering of the steering wheel perfor-mance or steering wheel returnability is better when thesteering torque gradient is large and the hysteresis is small [9]Based on this fact the result is shown in Figure 21(c) whereby
the hysteresis of the steering wheel returnability based onexperimental result is smaller compared to EPS steeringsystem In addition the steering torque gradient is larger thanEPS steering torque Therefore it is considered that the pro-posed feedback torque algorithmnot only achieved a steeringfeel but also improved the steering wheel returnability
46 Lower Speed Manoeuvre The steering torque results atlow speed manoeuvre are shown in Figures 22(b) and 22(c)when the input steering wheel is as shown in Figure 22(a)The vehicle cruised at constant speed of 10 kmh
The experimental steering torque is almost comparableto EPS torque but the amplitude is decreases The torque isdecreased due to the fact that the vehicle speed is reducedwhich is effect on reducing the demand to the total feedbacktorque Moreover the large turning of the steering wheelincreases the backlash of the motor which then contributes
International Journal of Vehicular Technology 15
2 3 4 5 6 7 8 9 10minus30
minus20
minus10
0
10
20
30
40
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle at 120 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steeringSimulationExperiment
10
5
0
minus5
minus10
3 4 5 6 7 8 9
(b) Steering torque estimation at 120 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment10
15
5
0
minus5
minus10
minus15minus20 minus10 0 10 20 30 40
Steering wheel angle (deg)
(c) Steering torque estimation against steering wheel angle at 120 kmh
Figure 21 Steering torque estimation response at 120 kmh
to the decrease of the motor torque Thus driver will sensesofter during steering the wheel during maneuver Howeverit has small different torque against EPS steering Wide rangeadjustment of gains in proposed control algorithm may helpimprove the steering torque response
3 4 5 6 7 8 9 10minus150
minus100
minus50
0
50
100
150
200
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
10
15
5
0
minus5
minus10
minus15
Time (s)2 3 4 5 6 7 8 9 10
(b) Steering torque estimation at 10 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
Steering wheel angle (deg)
10
8
6
4
2
0
minus2
minus4
minus6
minus8
minus10
minus100 minus50 0 50 100 150
(c) Steering torque estimation against steering wheel angle at 10 kmh
Figure 22 Steering torque estimation response at 10 kmh
5 Conclusion and Future Work
The proposed force feedback torque estimation and controlalgorithm have been described The proposed control algo-rithm is considered as an effort to generate the feedback
16 International Journal of Vehicular Technology
torque for driver steering feel in the vehicle SBW systemThe mathematical model of SBW subsystem consists ofsteering wheel and front axle systems that are modelledThe effectiveness of the control algorithm is verified throughthe experimental hardware in the loop (HIL) interfacedwith Matlab XPC target toolbox software According to theexperimental results the proposed control algorithm is ableto generate the feedback torque and the driver is able to sensesteering feel similar to that obtained in EPS steering systemThe LQR controller with gain scheduling based on steeringwheel angle and vehicle speed function provides better torquecontrol that allows rejection of the uncertainty torque froma road condition Furthermore the steering feel gain is ableto increase and lower the steering torque with improvedsteering stability at high speed and it improved the vehiclemanoeuvre at low speed It is also found that compensationtorque is able to improve the feedback torque and to stabilizethe system However it has small different torque magnitudethat occurs due to the backlash of steering wheel motor anduncertainty disturbance of mechanical properties Howeverwith a wide range adjustment of the gains on the proposedcontrol algorithm it may improve and provide realisticsteering torque for better driver steering feel
In future work the proposed control algorithm will betested in real vehicle in order to have more realistic steeringtorque for further investigation and analysis
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
Acknowledgments
This work is fully supported by the Ministry of Higher Edu-cation Malaysia and University Technology Malaysia underFRGS (vote 4F370) and Research University Grant (vote00G64) This work is also supported by Proton HoldingBerhad
References
[1] Y Yao ldquoVehicle steer-by-wire system controlrdquo SAE TechnicalPaper 2006-01-1175 2006
[2] E A Kumar D Dinesh and N Kamble ldquoAn overview of activefront steering systemrdquo International Journal of Scientific ampEngineering Research vol 3 no 6 2012
[3] S M H Fahami H Zamzuri S A Mazlan and N B Zulkar-nain ldquoThe design of vehicle active front steering based on steerby wire systemrdquo Advanced Science Letters vol 19 no 1 pp 61ndash65 2013
[4] S M H Fahami H Zamzuri S A Mazlan and S A SaruchildquoThe variable steering ratio for vehicle steer by wire systemusing hyperbolic tangent methodrdquo Applied Mechanics andMaterials vol 575 pp 781ndash784 2014
[5] T Kaufmann S Milsap B Murray and J Petrowski ldquoDevelop-ment experience with steer by wirerdquo in Proceedings of the SAEInternational Congress and Exhibition SAE 2001-01-2479 2001
[6] S Amberkare F Bolourchi J Demerly and S Millsap ldquoAcontrol system methodology for steer by wire systemsrdquo SAETechnical Paper 2004-01-1106 2004
[7] A Baviskar J R Wagner D M Dawson D Braganza andP Setlur ldquoAn adjustable steer-by-wire haptic-interface trackingcontroller for ground vehiclesrdquo IEEE Transactions on VehicularTechnology vol 58 no 2 pp 546ndash554 2009
[8] M Segawa S Kimura T Kada and S Nakona ldquoA study of reac-tive torque control for steer bywire systemrdquo inProceedings of theInternational Symposium on Advanced Vehicle Control (AVECrsquo02) pp 653ndash658 Hiroshima Japan September 2002
[9] S W Oh H C Chae S C Yun and C S Han ldquoThe design ofa controller for the steer-by-wire systemrdquo JSME InternationalJournal Series C Mechanical Systems Machine Elements andManufacturing vol 47 no 3 pp 896ndash907 2004
[10] C J Kim J H Jang S K Oh J Y Lee and J K HedrickldquoDevelopment of a control algorithm for a rack-actuating steer-by-wire system using road information feedbackrdquo Proceedingsof the Institution of Mechanical Engineers Part D Journal ofAutomobile Engineering vol 222 no 9 pp 1559ndash1571 2008
[11] S Asai H Kuroyanagi and S Takeuchi ldquoDevelopment of asteer-by-wire system with force feedback using a disturbanceobserverrdquo inProceedings of the SAEWorldCongressampExhibitionon Steering amp Suspension Technology Detroit Mich USA 2004
[12] T J Park C S Han and S H Lee ldquoDevelopment of the elec-tronic control unit for the rack-actuating steer-by-wire usingthe hardware-in-the-loop simulation systemrdquo Mechatronicsvol 15 no 8 pp 899ndash918 2005
[13] D Odenhtal T Bunte H D Heitzer and C Eicker ldquoHow tomake steer-by-wire feel like power steeringrdquo in Proceedings ofthe 15th IFACWorld Congress Barcelona Spain July 2000
[14] H Na C Zong and D Hu ldquoInvestigations on corneringcontrol algorithm design and road feeling optimization for asteer-by-wire vehiclerdquo in Proceedings of the IEEE InternationalConference on Mechatronics and Automation (ICMA rsquo09) pp3246ndash3251 Changchun China August 2009
[15] S Ancha A Baviskar J RWagner andDMDawson ldquoGroundvehicle steering systems modelling control and analysis ofhydraulic electric and steer-by-wire configurationsrdquo Interna-tional Journal of Vehicle Design vol 44 no 1-2 pp 188ndash2082007
[16] P Yih and J C Gerdes ldquoModification of vehicle handlingcharacteristics via steer-by-wirerdquo IEEE Transactions on ControlSystems Technology vol 13 no 6 pp 965ndash976 2005
[17] S M H Fahami H Zamzuri S A Mazlan and M A ZakarialdquoModeling and simulation of vehicle steer by wire systemrdquoin Proceedings of the IEEE Symposium on Humanities Scienceand Engineering Research (SHUSER rsquo12) pp 765ndash770 KualaLumpur Malaysia June 2012
[18] S Anwar and L Chen ldquoAn analytical redundancy-based faultdetection and isolation algorithm for a road-wheel controlsubsystem in a steer-by-wire systemrdquo IEEE Transactions onVehicular Technology vol 56 no 5 pp 2859ndash2869 2007
[19] M T Do Z Man C Zhang H Wang and F S Tay ldquoRobustsliding mode-based learning control for steer-by-wire systemsin modern vehiclesrdquo IEEE Transactions on Vehicular Technol-ogy vol 63 no 2 pp 580ndash590 2014
[20] J Duan R Wang and Y Yu ldquoResearch on control strategies ofsteer-by-wire systemrdquo in Proceedings of the International Con-ference on Intelligent Computation Technology and Automation(ICICTA rsquo10) vol 2 pp 1122ndash1125 May 2010
International Journal of Vehicular Technology 17
[21] B Zheng and S Anwar ldquoYaw stability control of a steer-by-wireequipped vehicle via active front wheel steeringrdquoMechatronicsvol 19 no 6 pp 799ndash804 2009
[22] S Sahoo S C Subramanian and S Srivastava ldquoDesign andimplementation of a controller for navigating an autonomousground vehiclerdquo in Proceedings of the 2nd International Confer-ence on Power Control and Embedded Systems (ICPCES rsquo12) pp1ndash6 IEEE Allahabad India December 2012
[23] T D Gillespie Fundamentals of Vehicle Dynamics Society ofAutomotive Engineers 1992
[24] M H Lee S K Ha J Y Choi and K S Yoon ldquoImprovement ofthe steering feel of an electric power steering system by torquemap modificationrdquo Journal of Mechanical Science and Technol-ogy vol 19 no 3 pp 792ndash801 2005
[25] R Pastorino M A Naya J A Perez and J Cuadrado ldquoGearedPM coreless motor modelling for driverrsquos force feedback insteer-by-wire systemsrdquo Mechatronics vol 21 no 6 pp 1043ndash1054 2011
[26] P Koehn andM Eckrich ldquoActive steeringmdashthe BMWapproachtowards modern steering technologyrdquo SAE Technical PaperSeries 2004-01-1105 SAE International 2004
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
International Journal of Vehicular Technology 3
120591totalfeedback
ks1
120575sw
bm1
120575m1
Jm1
L1
R1
Vs1
ia1
(a)
Vs2
L2
R2
ia2
bm2
120575m2
Jm2
ks2
brackf
Mrackf
yrackfklf klfBkpf Bkpf
Jm2 Jm2
(b)
Figure 2 System diagram (a) steering wheel and (b) front axle system [15]
Consequently the state equation of steering wheel system isgiven as
119904119898
= 119860119904119898
119909119904119898
+119861119904119898
119906119904119898
119910119904119898
= 119862119904119898
119909119904119898
119909119904119898
= [ 1205751198981 1205751198981 1198941198861]
(3)
And 119906119904119898
is considered the input of the steering wheel motorand the parameter states are
119860119904119898
=
[[[[[[
[
minus(11988711989811198691198981
) 01198961199041
1198691198981
1 0 0
minus(119896119904119898
1198711) 0 minus(
11987711198711
)
]]]]]]
]
119861119904119898
=
[[[[
[
0
0
11198711
]]]]
]
119906119904119898
= 120591total feedback = 1198811199041
(4)
22 Front Wheel System The function of the front axle sys-tem is to ensure the front tire angle follows the steering wheelangle command according to the steering ratio [17]The frontaxle is composed of a front axle DC motor rack and pinionand tire model The system diagram of the front axle systemand the subsystem block diagram are shown in Figures 2(b)and 3 respectively The input to the system is the steeringwheel angle (120575
119904119908) while the front tire angle (120575
119891) is considered
the output The parameters of the front axle system areapproximated assumption that are illustrated in Table 2
Table 2 Front axle system parameter [15]
Parameter Description Value Unit1198772 Motor resistance 464 ohm1198712 Motor inductance 0015 henry119870119891119898
Motor constant 0032 Nm1198691198982 Motor inertia 00062 kgm2
1198871198982 Motor damping 00036 Nm(rads)1198961199042 Lumped torque stiffness 0025 Nm
119861rack Rack damping coefficient 0015 henry119872rack Rack lumped mass 0032 Nm119896119897119891
Rack linkage stiffness 000062 kgm2
119903119871
Offset of king pin axis 000036 m119903119901
Pinion gear radius 0025 m119861119896119901
King pin damping coefficient 000062 kgm2
119868119891
Lumped front wheel inertia 000036 kgm2
1198961199042 Lumped torque stiffness 0025 Nm
120575sw 120575m2 120575fyrack
Front axle system
Front axleDC motor
Rack andpinionsystem
Front tiredynamics
Figure 3 Block diagram of front axle subsystem
The mathematical equation of the front axle DC motor iswritten as follows
The front axle motor angular displacement is
1205751198982 = minus
11988711989821198691198982
1205751198982 +
1198961199042
1198691198982
1198941198862 (5)
4 International Journal of Vehicular Technology
and the front axle motor current is
1198941198862 = minus
11987721198712
1198941198862 +
119896119891119898
1198712
1205751198982 +
1198811199042
1198712 (6)
Consequently the state equations of the front axle motor aregiven as
119891119898
= 119860119891119898
119909119891119898
+119861119891119898
119906119891119898
119910119891119898
= 119862119891119898
119909119891119898
119909119891119898
= [ 1205751198982 1205751198982 1198941198862]
(7)
And 119906119891119898
is the input to the front axle motor while outputis the rate change of front axle motor angular displacement( 1205751198982) the front axle motor angular displacement (120575
1198982) andthe current of front axle motor (119894
1198862)The parameter states are written as follows
119860119891119898
=
[[[[[[[
[
minus(11988711989821198691198982
) 01198961199042
1198691198982
1 0 0
minus(
119896119891119898
1198712) 0 minus(
11987721198712
)
]]]]]]]
]
119861119891119898
=
[[[[
[
0
0
11198712
]]]]
]
119906119891119898
= 120575119904119908
= 1198811199042
(8)
The rack and pinion system is
119910rack =1
119872rack[(minus
2119896119897119891119910rack
1199032119871
)minus(1198961199042119910rack1199032119901
)
minus119861rack119891 119910rack +(11989611990421205751198982119903119901
)]
(9)
And 119906119910rack is considered the input of the rack and pinion
system and the output is the rack displacement (119910rack) andthe parameter states are as follows
119860119910rack
=[[[
[
0 1
minus[(
2119896119897119891
119872rack1199032119871
) + (1198961199042
119872rack1199032119901
)] minus(119861rack119872rack
)
]]]
]
119861119910rack =
[[
[
0
(1198961199042
119872rack119903119901)
]]
]
119906119910rack = 120575
1198982
(10)
Fyr
lr lf
r
120572r
120573
Fyf120572f
120575f
Figure 4 Single track-linear vehicle model
and the front tire dynamic is
120575119891=
1119868119891
[(minus119896119897119891120575119891+(
119896119897119891119910rack
119903119871
)minus119861119896119901
120575119891)] (11)
and 119906tire is considered the input of the front tire dynamicTheoutput is the front tire angle (120575
119891) and the parameter state is
119860 tire =[[
[
0 1
minus(
119896119897119891
119868119891
) minus(
119861119896119901
119868119891
)
]]
]
119861tire =[[
[
0
(
119896119897119891
119868119891119903119871
)
]]
]
119906tire = 119910rack
(12)
23 Single Track-Linear Vehicle Model The stability charac-teristics of a steer-by-wire vehicle are affected by the influenceof front tires via the feedback of the vehicle dynamic responseThus it is necessary to develop a single track-linear vehiclemodel as mentioned by author [16 18ndash21] in simulation andexperimental validationThe single track linear vehiclemodelis shown in Figure 4 In this paper the aim is to model thelinear vehicle in order to generate the estimation of self align-ing torque (120591
119886) This acts as the disturbance input to the front
axle system and assures the road contact with front tire Thedynamics of single track-linear vehicle model is representedby bicycle model where the states contain yaw rate (119903) andvehicle body slip angle (120573)
The following assumptions are considered for normaldriving maneuverer These assumptions could simplify thevehicle model [18]
(i) friction force in the 119909-direction is negligible
(ii) vehicle speed is constant
(iii) angle for right and left tire is approximately the same
International Journal of Vehicular Technology 5
With the aforementioned assumptions the derived equationsof the motion of yaw rate (119903) and vehicle body slip angle (120573)are shown in (13) as follows
119903 = [
[
(1198972119891119862119891+ 119897
2119903119862119903)
119868119904119881
]
]
119903+[
[
(1198972119903119862119903minus 119897
2119891119862119891)
119868119904
]
]
120573
+[
(119897119891119862119891)
119868119904
]120575119891
120573 = [1+(119897119903119862119903minus 119897119891119862119891)
1198981198812 ] 119903 minus[
(119862119891+ 119862119903)
119898119881]120573
+[
119862119891
119898119881]120575119891
(13)
In the linear tire region lateral force at the front and rear isrelated to slip angle by cornering stiffness of the front and reartires written as follows
119865119910119891
= minus119862119891120572119891
119865119910119903
= minus119862119903120572119903
(14)
At a larger slip angle the relation between lateral force andslip angle becomes nonlinear By assuming a smaller slipangle approximation the slip angle is written in terms ofvehicle body slip angle and yaw rate that are considefed asfollows
120572119891= (120573+
119897119891
119881119903minus 120575119891)
120572119903= (120573+
119897119903
119881119903minus 120575119903)
(15)
Rewriting (13) into state space form we haveV119898 = 119860V119898119909V119898 +119861V119898119906V119898
119910V119898 = 119862V119898119909V119898
119909V119898 = [119903 120573]
(16)
And 119906V119898 is considered the input to the linear vehicle modelwhile the yaw rate (119903) and vehicle body slip angle (120573) arethe output of the linear vehicle model The parameter stateis written as follows119860V119898
=
[[[[[[
[
(
(1198712119891119862119891+ 119871
2119903119862119903)
119868119904119881
) minus
(119871119903119862119903minus 119871119891119862119891)
119868119904
1 + (
(119871119903119862119903minus 119871119891119862119891)
1198981198812 ) minus
(119862119891+ 119862119903)
119898119881
]]]]]]
]
119861V119898 =[[[
[
119871119891119862119891
119868119904
119862119891
119898119881
]]]
]
119906V119898 = 120575119891
(17)
The parameter of a linear vehicle model directly can bemeasured but some parameters need to be estimated Thereare several methods to estimate themodel parameter [22 23]A method based on Sahoo et al [22] is used to estimate themodel parameters with assumption that the vehicle mass isdistributed to front and rear of each tire The total of vehiclemass (119898) that is based on the front (119898
119891) and rear (119898
119903) tire
masses is written as follows
119898 = 119898119891+119898119903
(18)
when
119898119891= 119898119891119897+119898119891119903
119898119903= 119898119903119897+119898119903119903
(19)
where 119898119891119897 119898119891119903 119898119903119897 and 119898
119903119903are the masses of the vehicle
at front left front right rear left and rear right respectivelyThe parameter length of wheel base (119897) which is sum of lengthfrom a center of front tire 119897
119891and rear tire 119897
119903to COG of vehicle
body is written as follows
119897 = 119897119891+ 119897119903 (20)
while the inertia of vehicle 119868119904that is approximated by assum-
ing the vehicle as two point masses joined with mass-lessroad [22] is written as follows
119868119904= 1198981198911198972119891+1198981199031198972119903 (21)
The front cornering stiffness 119862119891and rear cornering stiffness
119862119903parameters are written in (22) by assuming that the type
of radial tire is used and cornering stiffness per degree of slipangle is approximately 16-17 of the load on the tire [22]Consider
119862119891= 119898119891lowast 119892 lowast 0165 (Ndeg)
119862119903= 119898119903lowast 119892 lowast 0165 (Ndeg)
(22)
The parameters of the linear vehicle model are illustrated inTable 3
3 Force Feedback Torque EstimationControl Algorithm
It has been known that the force feedback torque for driversteering feel should be created artificially due to the elimina-tion of mechanical column shaft in SBW system
Moreover the proportion of the steering feel obtainedshould be similar to conventional steering system For thisreason the steering wheel in SBW system that is equippedwith the motor actuator is used to generate the steeringfeel by controlling the total feedback torque (120591total feedback)A schematic block diagram of the proposed force feedbacktorque control is shown in Figure 5
Based on Figure 5 the total force feedback torque consistsof the torque of front axle motor (120591
119891119898) torque on steering
wheel motor (120591119904119898) and compensation torque which is the
inertia (120591inertia) and damping torque (120591damp) Meanwhile the
6 International Journal of Vehicular Technology
+
+
+
minus
minus
Wheel synchronization
120575sw
120575sw
120591a
Frontaxle
system
LQRcontroller
LQRcontroller
120575m2
120575m2
120575f
ia2
120573
r
Linearvehiclemodel
Force feedback torque
120591map
120591fm
120591sm
120591damp
120591inertia
VGain
scheduling
kfm
kfeelsum
120591total feedback Steeringwheelsystem
Steering wheel system
Steeringwheel
120575m1
120575m1
ia1
ksm
Estimation selfaligning torque
Torque(RMS)
Figure 5 Overview of force feedback torque control algorithm for SBW system
Table 3 Linear vehicle model [16]
Parameter Description Value Unit119862119891 Front cornering stiffness 40000 Nrad
119862119903 Rear cornering stiffness 35000 Nrad
119897119891
Length from center frontwheel to (COG) vehicle 14 m
119897119903
Length from center rearwheel to COG vehicle 1 m
119898 Mass of the vehicle 1535 kg119868119904 Inertia of vehicle 2149 kgm2
119881 Speed of vehicle mdash kmh119865119910119891 Front tire lateral force mdash Nm
119865119910119903 Rear tire lateral force mdash Nm
120572119891 Front tire slip angle mdash Nm
120572119903 Rear tire slip angle mdash Nm
torque map is based on the function of the steering wheelangle and vehicle speed is added to the total feedback torqueMoreover a steering feel gain (119896feel) is adapted based on thefunction of steering wheel angle and vehicle speed to varythe feedback torque In order to control the feedback torquethe gain scheduling with LQR controller is used To showthe influence of the tire on the road surface the self aligningtorque (120591
119886) acts as input disturbance to the front axle system
as follows
120591119886= minus119862
119891(119905119901+ 119905119898) [120573+
119897119891119903
119881minus 120575119891]120583 (23)
where 120583 is coefficient for road dry condition pneumatic trail(119905119901) and mechanical trail (119905
119898)
On the other hand any change of the road surface couldchange the feedback torque magnitude of the front axlemotor Furthermore the LQR controller of the front axle sys-tem is used to synchronize the front tire angle (120575
119891) with driver
input steering wheel angle (120575119904119908) by controlling the angle of
the front axle motor (1205751198982) in accordance with the steering
ratioThus based on the proposed control algorithm the total
force feedback torque to the steering wheel system is writtenas follows
120591total feedback = ([(radicmean (120591119904119898
+ 120591119891119898
+ 120591map))]
+ 120591inertia + 120591damp) 119896feel
(24)
Author [9 10] has generated and discussed the force feed-back torque using the torque map conceptThe torque map isbased on the function of the steering wheel angle and vehiclespeed parameter The feedback torque is determined by twocontrol parameters (119896
120573) for the vehicle speed and (119896
120572) for the
steering wheel angle functionThe torque based on the function of vehicle speed is given
as follows [9]
1205911 = 1198961205731199092(13119909minus
12119881max)+119879in (25)
International Journal of Vehicular Technology 7
0 2 4 6 8 10minus40
minus30
minus20
minus10
0
10
20
30
40
Time (s)
Stee
ring
torq
ue (N
m)
70kmh75kmh
50kmh40kmh30kmh20kmh10kmh
80kmh
60kmh
Figure 6 Steering torque based on torque map at different speed
where (119909) is initial velocity and (119881max) is maximum velocitywhile the torque based on a function of steering wheel angleis written as follows [9]
1205912 = 119896120572radic1003816100381610038161003816120575119904119908
1003816100381610038161003816 sgn 120575119904119908 (26)
Based on the torque map concept the results of steeringtorque at different vehicle speed are shown in Figure 6 whenthe input steering wheel angle is a double lane changemanoeuvreThe torque increased at high speed anddecreasedat low speed [9 24] By varying the two control parametersthe steering reactive torque changes at different torque mag-nitude However the feedback torque based on the torquemap concept did not create the artificial steering feel dueto the fact that no element of road contact such as vehicledynamic responses are feedback to the torque map modelThus driver not sense realistic driver steering feel
Thus for this reason the front axle motor torque andsteering wheel motor torque are added to the torque mapThese ensure that the driver senses the feedback torque fromroad contact The front axle motor torque is chosen due toits proportion to the road surface The reaction torque of theroad surface gives the effect to the rack and pinion systemand it connects to the front axle DC motor Thus the frontaxle motor torque would always be proportional to the roadsurface Any changes in road surface will influence the frontaxle motor torque In order to measure the torque at the frontaxle motor the direct currentmeasurement technique is usedto estimate the torque
The direct relationship between torque and current of amotor can be defined [25]Thus then front axlemotor torque(120591119891119898
) that consists of the front axle motor constant (119896119891119898
) andthe motor current (119894
1198862) is written as follows
120591119891119898
= 1198941198862119896119891119898 (27)
and the corresponding steeringwheelmotor torque (120591119904119898) that
consists of the steering wheel motor constant (119896119904119898) and the
motor current (1198941198861) is written as follows
120591119904119898
= 1198941198861119896119904119898 (28)
0 2 4 6 8 10Time (s)
Mot
or to
rque
(Nm
)
Front axle motor torque
Steering wheel motor torque6
4
2
0
minus2
minus4
minus6
Figure 7 Steering wheel and front axle motor torque at 80 kmh
Figure 7 shows the front axle motor and steering wheelmotor torque response when the input is a double lanechangemanoeuvreThe results show that the front axlemotorgives a high torque This is due to the fact that the inputdisturbance torque between the tire and road surface which isthe self aligning torque has an effect on the front axle motorIncreasing the vehicle speedwill increase the front axlemotortorque On the other hand the magnitude of steering wheelmotor torque is decreased due to the elimination of mechan-ical column shaft and is not affected by the self aligningtorque
The phase compensation torque consists of the inertiatorque and damping torque potential to recreate the artificialsteering feel of the conventional steering system Moreoverit is intended to adjust the steering feel reduce the vibrationand stabilize the system Due to this fact the element of com-pensation torque is taken in order to improve the feedbacktorque for a driver steering feel and to stabilize the system
The damping torque (120591damp) which is the rate change ofthe steering wheel motor angle ( 120575
1198981) and the damping gain(119896damp) is written as follows
120591damp = 1205751198981119896damp (29)
The inertia torque (120591inertia) that is the acceleration of thesteering wheel motor angle ( 120575
1198981) and the inertia gain (119896inertia)is written as follows
120591inertia = 1205751198981119896inertia (30)
The inertia and damping torque responses are shown inFigure 8 when the input steering wheel angle is a double lanechange manoeuvre It is obtained that the damping torquegives a smooth signal and reduces the vibration Meanwhilethe inertia torque gives high magnitude Thus the combina-tion between the inertia torque and damping torque not onlyprovides an average of steering feel but is also able to stabilizethe system at the same time Figure 9 shows the feedback
8 International Journal of Vehicular Technology
0 2 4 6 8 10minus6
minus4
minus2
0
2
4
6
Time (s)
Torq
ue (N
m)
Inertia torque Damping torque
Figure 8 Inertia and damping torque
0 2 4 6 8 10Time (s)
Forc
e fee
dbac
k to
rque
(Nm
)
15
10
5
0
minus5
minus10
minus15
minus20
minus25
10kmh
50kmh100 kmh120 kmh
135kmh
Figure 9 Proposed force feedback torque response at differentvehicle speed
torque response based on (24) The torque is increasedaccording to the vehicle speed The higher vehicle speed willincrease the feedback torque and vice versa due to the samefact that is proposed by torque map
31 Linear Quadratic Regulator (LQR) Controller The LQRcontroller provides the best possible performance to controland stabilize the system by changing the location of the polesof the system to the optimal location for a time responseovershoot and steady stateThis is done by designing the statefeedback control119870 thatminimized cost function 119869 to stabilizethe system In this paper the LQR controller is used firstlyto control the front axle motor angle for the wheel synchro-nization and secondly to control the steering wheel motorfor a feedback torque
The wheel synchronization or directional control is abasic characteristic for steering function where a front tire
+
minus
sum
k1
k2
k3
x1
x2
x3
Plantr(t)
LQR controller
Figure 10 Basic control block diagram of LQR controller
angle should follow driver input steering wheel angle accord-ing to the steering ratio This is done by controlling theangle of the front axle motor In the LQR control schemesa feedback gain matrix is designed to achieve some compro-mise between the use of control effort the magnitude andthe speed of response that will guarantee a stable systemFigure 10 shows the control block diagram of the LQRcontroller
The optimal LQRproblem is often definedmore generallyand consists of finding controller transfer matrix that mini-mizes the cost function 119869 that is written as follows
119869 = int
prop
0(119909119879
119876119909+119906119879
119877119906) 119889119905 (31)
where 119909119879 and119880119879 respectively represent the transpose of the
state and input vector The 119876 and 119877 are positive definite (orpositive-semi definite) real symmetric matrix The 119876 and 119877
are usually chosen to be diagonal matrix written as follows
119876119894119894=
1maximum acceptable value of 1199092
119894
119877119895119895
=1
maximum acceptable value of 1199062119895
(32)
where
119894 = 120576 1 2 ℓ
119895 = 120576 1 2 119895 (33)
In essence Brysonrsquos rule scales the variables that appear in 119869 sothat the maximum acceptable value for each term is oneThisis especially important when the units used for the differentcomponents of 119906 and 119909 make the values for these variablesnumerically very different from each other
Assume that the steering ratio is 15 1 and the gains ofthe LQR controller are [01 098 0008] Figure 11 shows aresponse of the front axle system when the input steeringwheel angle is a step function as shown in Figure 11(a)
The results show that the angle of the front axle motorhas a fast response compared to the EPS steering system
International Journal of Vehicular Technology 9
0 2 4 6 8 100
2
4
6
8
10
12
14
16
18
20
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle
0 2 4 6 8 100
2
4
6
8
10
12
14
16
18
Time (s)
Mot
orc
olum
n an
gle (
deg)
EPS steeringmdashcolumn shaft angle
SBW systemmdashmotor angle
(b) Front axle motor angle
0 2 4 6 8 10
Rack
disp
lace
men
t (m
)
SBW system
EPS steering
Time (s)
18
16
14
12
10
8
6
4
2
0
times10minus3
(c) Rack displacement
0 2 4 6 8 100
02
04
06
08
1
12
14Fr
ont t
ire an
gle (
deg)
SBW system
EPS steering
Time (s)
(d) Front tire angle
Figure 11 Front axle system response for wheel synchronization
Moreover the LQR controller is able to reduce the delaycaused by the self aligning torque created from the contactbetween the tire and road surfaces This not only improvedthe vehicle stability but also ensured that the driver has amore confident level during manoeuvre Compared to theEPS steering system the delay caused by mechanical columnshaft connected to the steering wheel system Furthermore itcan be seen that the front tire angle is at almost fairly close to1∘ as shown in Figure 11(d) However it has small chatteringdue to the self aligning torque However with wide rangeadjustment gains of the LQR controller the steering responsecould be improved
32 Gain Scheduling (GS) with LQR Controller In certainsituations the dynamic behaviour of the process changed inaccordance with the conditions of the process This situation
could affect the stability of the system if not to control prop-erly Thus it is possible to change the gains of the controllerby monitoring the operating conditions of the process tohave a stable system The gain scheduling (GS) techniqueis used to change the gain of the LQR controller in order tocontrol the feedback torque and the steering feel gain (119896feel)is used to increase and decrease the torque It is important toincrease the feedback torque at high speed for steering stabil-ity and to decrease it in low speed especially during parkingto improve vehicle manoeuvre [9] This is based on the steer-ingwheel angle and vehicle speed function in order to achievethe feedback torque for the real driver steering feel
The relationship between a gain scheduling with LQRcontroller is shown in Table 4 The vehicle speed is dividedinto 3 categories at low speed (0ndash30 kmh) medium speed(31ndash100 kmh) and high speed (101ndash120 kmh) The maxi-mum vehicle speed is limited to 120 kmhThe steering wheel
10 International Journal of Vehicular Technology
0 2 4 6 8 10minus50
minus40
minus30
minus20
minus10
0
10
20
30
40
50
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Input steering wheel angle
0 1 2 3 4 5 6 7 8 9 10minus15
minus10
minus5
0
5
10
Time (s)
Stee
ring
torq
ue (N
m) LQR
LQR + GS
Reference
(b) 120583 = 1
0 1 2 3 4 5 6 7 8 9 10
minus15
minus10
minus5
0
5
10
15
Stee
ring
torq
ue (N
m) LQR
Reference
Time (s)
LQR + GS
(c) 120583 = 14
0 1 2 3 4 5 6 7 8 9 10minus20
minus15
minus10
minus5
0
5
10
15
20
Stee
ring
torq
ue (N
m) LQR LQR + GS
Reference
Time (s)
(d) 120583 = 17
Figure 12 Comparison of torque response of LQR and LQR + GS controller at different 120583
Table 4 The gains of LQR controller and steering feel
119881 (kmh) 120575119904119908
(degree) 1198961 1198962 1198963 119896feel
(0ndash30) (plusmn180∘ lt 120575119904119908
lt plusmn405∘) 003 065 0001 014(31ndash100) (plusmn46∘ lt 120575
119904119908lt plusmn179∘) 003 084 0001 060
(101ndash120) 120575119904119908
le +45∘ or 120575119904119908
le minus45∘ 005 095 0001 095
angle is also divided into 3 categories according to vehiclespeed condition The notation ldquo+rdquo means that the driverturned the steering wheel clockwise and notation ldquominusrdquo is usedfor counter clockwise in which the maximum steering wheelturn is plusmn405∘
The gains of the LQR controller are defined using Brysonrsquosrule method and the steering feel gain is defined usingintuitive and has a range from 01 to 1 This range has beendefined for various speeds 10 kmh 80 kmh and 120 kmhafter comparing with EPS steering system [15] For exampleif the driver turns a steering wheel between (plusmn180∘ lt 120575
119904119908lt
plusmn405∘) and vehicle speed is 10 kmh the total feedback torque
is multiplied by steering feel gain 119896feel = 014 Thus thefeedback torque is decreased which results in making thedriver feel it easier turning the steering wheel and improvesvehicle manoeuvre The feedback torque is regulated by LQRcontroller when the gains are 1198961 = 003 1198962 = 065 and1198963 = 0001 The gains are the optimum that has been assuredby providing a better torque control based on vehicle speedand steering wheel angle function
Comparing torque response between LQR and LQR +GScontroller for different coefficient (120583) is shown in Figure 12when the input steering wheel angle is a double lane changemanoeuvre as shown in Figure 12(a) The vehicle speed isset to be 80 kmh Based on results obtained the LQR + GStechnique offers an improvement 90 and the curve trend isalmost similar to reference torque compared to the LQR con-troller This is because the gains of LQR controller changedbased on the steering wheel angle and vehicle speed param-eter which is able to reject the uncertainty torque of a roadcondition Meanwhile with the fixed gains of the LQR con-troller it has an excessive torque that can cause the system tobe unstable The RMS torque values are illustrated in Table 5
International Journal of Vehicular Technology 11
Table 5 Comparison of RMS torque values of LQR and LQR + GScontroller
Reference LQR LQR + GS120583 = 1 56361 50795 55450120583 = 14 61218 68339 57243120583 = 17 66184 72888 62080
minus50 0 50minus15
minus10
minus5
0
5
10
15
Steering angle (deg)
Stee
ring
torq
ue (N
m) Feel gain = 5
Feel gain = 01Feel gain = 05
Feel gain = 1
(a) Different steering feel gain at 80 kmh
minus50 minus40 minus30 minus20 minus10 0 10 20 30 40 50
minus15
minus10
minus5
0
5
10
15
Steering wheel angle (deg)
Stee
ring
torq
ue (N
m)
80kmh kfeel = 060
120 kmh kfeel = 095
10kmh kfeel = 014
(b) Different steering feel gain at various speed
Figure 13 Steering feel gains response at various vehicle speed
Figure 13(a) shows the steering torque with respect tosteering wheel angle when different steering feel gain is usedat various vehicle speeds It is observed that when the steeringfeel gain is increased at maximum range which is 1 thesteering torque will increase and driver will sense stiffer onthe steering wheel at high speed manoeuvre Meanwhilewhen decreasing the steering feel gain this will result inreducing the steering torqueThus the driver will sense softer
Target monitor
Host PC
Cross over cableConnector block
PCI 6221
Power supplyCurrent sensor
DC motor driver
Figure 14 Steering wheel system hardware overview
during turning the steering wheel at the low speed manoeu-vre However when the steering feel gain is 5 which exceedsits maximum limit the steering torque becomes unstableFigure 13(b) shows the steering feel gains at various speedsAt high speed of 120 kmh the steering feel gain is increasedand this will increase the steering torque for steering stabilityMeanwhile at low speed the steering feel gain is decreasedwhich will help improve vehicle manoeuvre especially duringparking Based on the proposed gain scheduling with theLQR controller not only does it provide better control on thesteering torque but also it is optional for a designer to have awide range adjustment to vary the steering torque for a betterdriver steering feel and to improve vehicle manoeuvre
4 Hardware in the Loop (HIL) and Result
For hardware in the loop (HIL) setup the front axle systemand linear vehicle model are simulated on the host computerThe steering wheel system is replaced by HIL mechanismThe position of the steering wheel motor is measured usinga rotary encoder with 500 pulses per revolution To measurethe steering wheel torque the current sensor is connectedin series with the DC motor driver The steering wheel DCmotor has a torque 129Nm with 100 rpm The single tracklinear vehicle model is used to generate the estimation selfaligning torque as element to the feedback torque and distur-bance to the front axle system The HIL system is executedunder 1ms sampling time and interfaced using Matlab-XPCtarget toolbox configuration as shown in Figures 14 and 15
The assumptions through the experiment are listed asfollows
(i) the vehicle cruises at constant speed(ii) the condition of dry asphalt road (120583 = 085)
The experiments are done under high medium and lowspeed manoeuvre [26] The same input steering wheel anglebetween simulation and experiment is given Therefore theresult obtained from real time control is compared with theelectronic power steering (EPS) system [15]
41 Motor Model Validation for Steering Wheel SystemAlthough the model of steering wheel system has been
12 International Journal of Vehicular Technology
Stee
ring
whe
el sy
stem
DC motor
Current
Encoder
16 bit card PCI 6221
Analogoutput
Analoginput
Counterchannel
TargetPC
Monitortarget
PC
Crossover cable
Host PC
Front axle systemand
linear vehicle model
Hardware part Simulation part
Figure 15 Block diagram of the steering wheel-XPC target interfaced
Steering wheel model
120575m1
120575m1
ia1
r(t)
Figure 16 Block diagrammotormodel validation for steeringwheelsystem
derived the model validation is necessary to ensure thesystem is fairly able to adequate for real time applicationThis section describes the motor model validation of steeringwheel system The block diagram is shown in Figure 16
The input signal is 119903(119905) and the output is rate change ofmotor ( 120575
1198981) motor position (1205751198981) and the motor current
( 1198941198981) The measured output response is then used and com-
pared with simulation to estimate the parameter of the steer-ing wheel model using the Matlab tool software The mea-sured output data are taken based on the open loop systemwhen the input to the system is sinusoidal signal withamplitude of plusmn45∘ for 10 sec The Gaussian method is usedto estimate the parameters of the steering wheel motorThe comparison between simulation and measured outputis shown in Figure 17 It is obtained that the output of thesimulation is almost identical to the measured output How-ever small difference errors occur due to the small parametricuncertainties that have not been modelled The parameterestimation value is illustrated in Table 6
42 Noise Filtering Both analog and digital devices have atrait of noise or unwanted features that have effect on thesystem performance such as electromagnetic interferenceThe noise could be either random or white noise Attenuationof this noise is often as a primary goal in control systemapplication and in particular when controlling the processsystem The average value of noise typically is zero whichwill give misreading information to the process and it wouldnot be possible to control the process In order to reduce thenoise inference filtering technique is used A filter is used toremove unwanted noise It removes some noise frequenciesto suppress interfacing signal and reduce background noise
Table 6 Parameter estimation for DCmotor steering wheel system
Parameter Value Estimation value Unit1198771 4405 564 ohm1198711 0010 0017 henry119870119904119898
1459 0024 Nm1198691198981 0002 00036 (kgm2)1198871198981 0004 00068 Nm(rads)1198961199041 165 0025 Nm
There are many types of filtering techniques such as low passfilter that has been used in this study The low pass filter isused to filter out the interference noise from the output of thecurrent sensor
Figure 18 shows the comparison of the output responsefor a current sensor with and without using filters Withouta filter the current sensor has high amplitude interferencenoise However with low pass filter under the frequency of5Hz the interference noise is reduced and the fluctuations areremoved Thus this provides a smoother output response
43 Effect on Compensation Torque A steering torque com-parison without and with compensation torque is shown inFigure 19The results for without compensation torque showsthat the steering torque response curve is far away fromEPS steering system with different amplitude as shown inFigure 19(a) This happens due to less of derivative reactiontorque from the steering wheel motor Figure 19(b) showssteering torque response with compensation torque
It is observed that by adding the additional compensationtorque the steering torque response curve trends fairly closeto EPS steering system Thus it helps to improve the driversteering feel during manoeuvre and stabilize the systemHowever with the wide range adjustment gains of the com-pensation torque it could improve the steering torqueresponse
44 Medium Speed Manoeuvre The experimental steeringtorque results at medium speed manoeuvre are shown inFigure 20(b) when the steering wheel angle is plusmn100∘ as shown
International Journal of Vehicular Technology 13
0 2 4 6 8 10minus50
0
50
Time (s)
Rate
chan
ge m
otor
pos
ition
(deg
s)
MeasurementSimulation
(a) Rate change of motor position
0 2 4 6 8 100
10
20
30
40
50
60
Time (s)
Mot
or p
ositi
on (d
eg)
Measurement
Simulation
(b) Position of the motor
0 2 4 6 8 10minus01
minus005
0
005
01
015
Mot
or cu
rren
t (A
)
MeasurementSimulation
Time (s)
(c) Current of the motor
Figure 17 Motor model validation comparison between simulation and measurement
With filter Without filter
Time (s)
Curr
ent (
A)
008
006
004
002
0
minus002
minus004
minus006
minus0081 15 2 25 3 35 4 45 5 55
Figure 18 Noise filtering using low pass filter
in Figure 20(a) From the results the experimental steeringtorque response is almost identical to the EPS steering systemHowever it has small different torque amplitudes These
occur due to the fact that mechanism or friction from themechanical gearbox motor is not modelled in proposedcontrol algorithm
14 International Journal of Vehicular Technology
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
minus152 3 4 5 6 7 8 9 10
(a) Without compensation torque
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
21 3 4 5 6 7 8 9 10
(b) With compensation torque
Figure 19 Steering torque influence on compensation torque
minus80
minus60
minus40
minus20
0
20
40
60
80
100
Stee
ring
whe
el an
gle (
deg)
Time (s)3 4 5 6 7 8 9 10
(a) Steering wheel angle at 80 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
15
20
5
0
minus5
minus10
minus15
minus203 4 5 6 7 8 9 10
(b) Steering torque estimation at 80 kmh
Figure 20 Steering torque estimation and steering wheel angle response at 80 kmh
45 High Speed Manoeuvre At high speed manoeuvre theinput steering wheel angle that is more than plusmn45∘ is very dan-gerous [12] Therefore in this experiment the input steeringwheel angle is less than plusmn45∘ as shown in Figure 21(a) Basedon the results shown in Figures 21(b) and 21(c) the experi-mental steering torque is obtained fairly similar to the EPSsteering system Although the steering wheel angle is at smallangles the influence of vehicle speed cause the steering torqueis increased Moreover at higher speed the vehicle has moreunder steer propensity [9] This increased the self aligningtorque in proportional high torque at front axle motorThusthe driver will sense stiffer on the steering wheel duringmanoeuvre
In general the centering of the steering wheel perfor-mance or steering wheel returnability is better when thesteering torque gradient is large and the hysteresis is small [9]Based on this fact the result is shown in Figure 21(c) whereby
the hysteresis of the steering wheel returnability based onexperimental result is smaller compared to EPS steeringsystem In addition the steering torque gradient is larger thanEPS steering torque Therefore it is considered that the pro-posed feedback torque algorithmnot only achieved a steeringfeel but also improved the steering wheel returnability
46 Lower Speed Manoeuvre The steering torque results atlow speed manoeuvre are shown in Figures 22(b) and 22(c)when the input steering wheel is as shown in Figure 22(a)The vehicle cruised at constant speed of 10 kmh
The experimental steering torque is almost comparableto EPS torque but the amplitude is decreases The torque isdecreased due to the fact that the vehicle speed is reducedwhich is effect on reducing the demand to the total feedbacktorque Moreover the large turning of the steering wheelincreases the backlash of the motor which then contributes
International Journal of Vehicular Technology 15
2 3 4 5 6 7 8 9 10minus30
minus20
minus10
0
10
20
30
40
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle at 120 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steeringSimulationExperiment
10
5
0
minus5
minus10
3 4 5 6 7 8 9
(b) Steering torque estimation at 120 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment10
15
5
0
minus5
minus10
minus15minus20 minus10 0 10 20 30 40
Steering wheel angle (deg)
(c) Steering torque estimation against steering wheel angle at 120 kmh
Figure 21 Steering torque estimation response at 120 kmh
to the decrease of the motor torque Thus driver will sensesofter during steering the wheel during maneuver Howeverit has small different torque against EPS steering Wide rangeadjustment of gains in proposed control algorithm may helpimprove the steering torque response
3 4 5 6 7 8 9 10minus150
minus100
minus50
0
50
100
150
200
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
10
15
5
0
minus5
minus10
minus15
Time (s)2 3 4 5 6 7 8 9 10
(b) Steering torque estimation at 10 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
Steering wheel angle (deg)
10
8
6
4
2
0
minus2
minus4
minus6
minus8
minus10
minus100 minus50 0 50 100 150
(c) Steering torque estimation against steering wheel angle at 10 kmh
Figure 22 Steering torque estimation response at 10 kmh
5 Conclusion and Future Work
The proposed force feedback torque estimation and controlalgorithm have been described The proposed control algo-rithm is considered as an effort to generate the feedback
16 International Journal of Vehicular Technology
torque for driver steering feel in the vehicle SBW systemThe mathematical model of SBW subsystem consists ofsteering wheel and front axle systems that are modelledThe effectiveness of the control algorithm is verified throughthe experimental hardware in the loop (HIL) interfacedwith Matlab XPC target toolbox software According to theexperimental results the proposed control algorithm is ableto generate the feedback torque and the driver is able to sensesteering feel similar to that obtained in EPS steering systemThe LQR controller with gain scheduling based on steeringwheel angle and vehicle speed function provides better torquecontrol that allows rejection of the uncertainty torque froma road condition Furthermore the steering feel gain is ableto increase and lower the steering torque with improvedsteering stability at high speed and it improved the vehiclemanoeuvre at low speed It is also found that compensationtorque is able to improve the feedback torque and to stabilizethe system However it has small different torque magnitudethat occurs due to the backlash of steering wheel motor anduncertainty disturbance of mechanical properties Howeverwith a wide range adjustment of the gains on the proposedcontrol algorithm it may improve and provide realisticsteering torque for better driver steering feel
In future work the proposed control algorithm will betested in real vehicle in order to have more realistic steeringtorque for further investigation and analysis
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
Acknowledgments
This work is fully supported by the Ministry of Higher Edu-cation Malaysia and University Technology Malaysia underFRGS (vote 4F370) and Research University Grant (vote00G64) This work is also supported by Proton HoldingBerhad
References
[1] Y Yao ldquoVehicle steer-by-wire system controlrdquo SAE TechnicalPaper 2006-01-1175 2006
[2] E A Kumar D Dinesh and N Kamble ldquoAn overview of activefront steering systemrdquo International Journal of Scientific ampEngineering Research vol 3 no 6 2012
[3] S M H Fahami H Zamzuri S A Mazlan and N B Zulkar-nain ldquoThe design of vehicle active front steering based on steerby wire systemrdquo Advanced Science Letters vol 19 no 1 pp 61ndash65 2013
[4] S M H Fahami H Zamzuri S A Mazlan and S A SaruchildquoThe variable steering ratio for vehicle steer by wire systemusing hyperbolic tangent methodrdquo Applied Mechanics andMaterials vol 575 pp 781ndash784 2014
[5] T Kaufmann S Milsap B Murray and J Petrowski ldquoDevelop-ment experience with steer by wirerdquo in Proceedings of the SAEInternational Congress and Exhibition SAE 2001-01-2479 2001
[6] S Amberkare F Bolourchi J Demerly and S Millsap ldquoAcontrol system methodology for steer by wire systemsrdquo SAETechnical Paper 2004-01-1106 2004
[7] A Baviskar J R Wagner D M Dawson D Braganza andP Setlur ldquoAn adjustable steer-by-wire haptic-interface trackingcontroller for ground vehiclesrdquo IEEE Transactions on VehicularTechnology vol 58 no 2 pp 546ndash554 2009
[8] M Segawa S Kimura T Kada and S Nakona ldquoA study of reac-tive torque control for steer bywire systemrdquo inProceedings of theInternational Symposium on Advanced Vehicle Control (AVECrsquo02) pp 653ndash658 Hiroshima Japan September 2002
[9] S W Oh H C Chae S C Yun and C S Han ldquoThe design ofa controller for the steer-by-wire systemrdquo JSME InternationalJournal Series C Mechanical Systems Machine Elements andManufacturing vol 47 no 3 pp 896ndash907 2004
[10] C J Kim J H Jang S K Oh J Y Lee and J K HedrickldquoDevelopment of a control algorithm for a rack-actuating steer-by-wire system using road information feedbackrdquo Proceedingsof the Institution of Mechanical Engineers Part D Journal ofAutomobile Engineering vol 222 no 9 pp 1559ndash1571 2008
[11] S Asai H Kuroyanagi and S Takeuchi ldquoDevelopment of asteer-by-wire system with force feedback using a disturbanceobserverrdquo inProceedings of the SAEWorldCongressampExhibitionon Steering amp Suspension Technology Detroit Mich USA 2004
[12] T J Park C S Han and S H Lee ldquoDevelopment of the elec-tronic control unit for the rack-actuating steer-by-wire usingthe hardware-in-the-loop simulation systemrdquo Mechatronicsvol 15 no 8 pp 899ndash918 2005
[13] D Odenhtal T Bunte H D Heitzer and C Eicker ldquoHow tomake steer-by-wire feel like power steeringrdquo in Proceedings ofthe 15th IFACWorld Congress Barcelona Spain July 2000
[14] H Na C Zong and D Hu ldquoInvestigations on corneringcontrol algorithm design and road feeling optimization for asteer-by-wire vehiclerdquo in Proceedings of the IEEE InternationalConference on Mechatronics and Automation (ICMA rsquo09) pp3246ndash3251 Changchun China August 2009
[15] S Ancha A Baviskar J RWagner andDMDawson ldquoGroundvehicle steering systems modelling control and analysis ofhydraulic electric and steer-by-wire configurationsrdquo Interna-tional Journal of Vehicle Design vol 44 no 1-2 pp 188ndash2082007
[16] P Yih and J C Gerdes ldquoModification of vehicle handlingcharacteristics via steer-by-wirerdquo IEEE Transactions on ControlSystems Technology vol 13 no 6 pp 965ndash976 2005
[17] S M H Fahami H Zamzuri S A Mazlan and M A ZakarialdquoModeling and simulation of vehicle steer by wire systemrdquoin Proceedings of the IEEE Symposium on Humanities Scienceand Engineering Research (SHUSER rsquo12) pp 765ndash770 KualaLumpur Malaysia June 2012
[18] S Anwar and L Chen ldquoAn analytical redundancy-based faultdetection and isolation algorithm for a road-wheel controlsubsystem in a steer-by-wire systemrdquo IEEE Transactions onVehicular Technology vol 56 no 5 pp 2859ndash2869 2007
[19] M T Do Z Man C Zhang H Wang and F S Tay ldquoRobustsliding mode-based learning control for steer-by-wire systemsin modern vehiclesrdquo IEEE Transactions on Vehicular Technol-ogy vol 63 no 2 pp 580ndash590 2014
[20] J Duan R Wang and Y Yu ldquoResearch on control strategies ofsteer-by-wire systemrdquo in Proceedings of the International Con-ference on Intelligent Computation Technology and Automation(ICICTA rsquo10) vol 2 pp 1122ndash1125 May 2010
International Journal of Vehicular Technology 17
[21] B Zheng and S Anwar ldquoYaw stability control of a steer-by-wireequipped vehicle via active front wheel steeringrdquoMechatronicsvol 19 no 6 pp 799ndash804 2009
[22] S Sahoo S C Subramanian and S Srivastava ldquoDesign andimplementation of a controller for navigating an autonomousground vehiclerdquo in Proceedings of the 2nd International Confer-ence on Power Control and Embedded Systems (ICPCES rsquo12) pp1ndash6 IEEE Allahabad India December 2012
[23] T D Gillespie Fundamentals of Vehicle Dynamics Society ofAutomotive Engineers 1992
[24] M H Lee S K Ha J Y Choi and K S Yoon ldquoImprovement ofthe steering feel of an electric power steering system by torquemap modificationrdquo Journal of Mechanical Science and Technol-ogy vol 19 no 3 pp 792ndash801 2005
[25] R Pastorino M A Naya J A Perez and J Cuadrado ldquoGearedPM coreless motor modelling for driverrsquos force feedback insteer-by-wire systemsrdquo Mechatronics vol 21 no 6 pp 1043ndash1054 2011
[26] P Koehn andM Eckrich ldquoActive steeringmdashthe BMWapproachtowards modern steering technologyrdquo SAE Technical PaperSeries 2004-01-1105 SAE International 2004
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
4 International Journal of Vehicular Technology
and the front axle motor current is
1198941198862 = minus
11987721198712
1198941198862 +
119896119891119898
1198712
1205751198982 +
1198811199042
1198712 (6)
Consequently the state equations of the front axle motor aregiven as
119891119898
= 119860119891119898
119909119891119898
+119861119891119898
119906119891119898
119910119891119898
= 119862119891119898
119909119891119898
119909119891119898
= [ 1205751198982 1205751198982 1198941198862]
(7)
And 119906119891119898
is the input to the front axle motor while outputis the rate change of front axle motor angular displacement( 1205751198982) the front axle motor angular displacement (120575
1198982) andthe current of front axle motor (119894
1198862)The parameter states are written as follows
119860119891119898
=
[[[[[[[
[
minus(11988711989821198691198982
) 01198961199042
1198691198982
1 0 0
minus(
119896119891119898
1198712) 0 minus(
11987721198712
)
]]]]]]]
]
119861119891119898
=
[[[[
[
0
0
11198712
]]]]
]
119906119891119898
= 120575119904119908
= 1198811199042
(8)
The rack and pinion system is
119910rack =1
119872rack[(minus
2119896119897119891119910rack
1199032119871
)minus(1198961199042119910rack1199032119901
)
minus119861rack119891 119910rack +(11989611990421205751198982119903119901
)]
(9)
And 119906119910rack is considered the input of the rack and pinion
system and the output is the rack displacement (119910rack) andthe parameter states are as follows
119860119910rack
=[[[
[
0 1
minus[(
2119896119897119891
119872rack1199032119871
) + (1198961199042
119872rack1199032119901
)] minus(119861rack119872rack
)
]]]
]
119861119910rack =
[[
[
0
(1198961199042
119872rack119903119901)
]]
]
119906119910rack = 120575
1198982
(10)
Fyr
lr lf
r
120572r
120573
Fyf120572f
120575f
Figure 4 Single track-linear vehicle model
and the front tire dynamic is
120575119891=
1119868119891
[(minus119896119897119891120575119891+(
119896119897119891119910rack
119903119871
)minus119861119896119901
120575119891)] (11)
and 119906tire is considered the input of the front tire dynamicTheoutput is the front tire angle (120575
119891) and the parameter state is
119860 tire =[[
[
0 1
minus(
119896119897119891
119868119891
) minus(
119861119896119901
119868119891
)
]]
]
119861tire =[[
[
0
(
119896119897119891
119868119891119903119871
)
]]
]
119906tire = 119910rack
(12)
23 Single Track-Linear Vehicle Model The stability charac-teristics of a steer-by-wire vehicle are affected by the influenceof front tires via the feedback of the vehicle dynamic responseThus it is necessary to develop a single track-linear vehiclemodel as mentioned by author [16 18ndash21] in simulation andexperimental validationThe single track linear vehiclemodelis shown in Figure 4 In this paper the aim is to model thelinear vehicle in order to generate the estimation of self align-ing torque (120591
119886) This acts as the disturbance input to the front
axle system and assures the road contact with front tire Thedynamics of single track-linear vehicle model is representedby bicycle model where the states contain yaw rate (119903) andvehicle body slip angle (120573)
The following assumptions are considered for normaldriving maneuverer These assumptions could simplify thevehicle model [18]
(i) friction force in the 119909-direction is negligible
(ii) vehicle speed is constant
(iii) angle for right and left tire is approximately the same
International Journal of Vehicular Technology 5
With the aforementioned assumptions the derived equationsof the motion of yaw rate (119903) and vehicle body slip angle (120573)are shown in (13) as follows
119903 = [
[
(1198972119891119862119891+ 119897
2119903119862119903)
119868119904119881
]
]
119903+[
[
(1198972119903119862119903minus 119897
2119891119862119891)
119868119904
]
]
120573
+[
(119897119891119862119891)
119868119904
]120575119891
120573 = [1+(119897119903119862119903minus 119897119891119862119891)
1198981198812 ] 119903 minus[
(119862119891+ 119862119903)
119898119881]120573
+[
119862119891
119898119881]120575119891
(13)
In the linear tire region lateral force at the front and rear isrelated to slip angle by cornering stiffness of the front and reartires written as follows
119865119910119891
= minus119862119891120572119891
119865119910119903
= minus119862119903120572119903
(14)
At a larger slip angle the relation between lateral force andslip angle becomes nonlinear By assuming a smaller slipangle approximation the slip angle is written in terms ofvehicle body slip angle and yaw rate that are considefed asfollows
120572119891= (120573+
119897119891
119881119903minus 120575119891)
120572119903= (120573+
119897119903
119881119903minus 120575119903)
(15)
Rewriting (13) into state space form we haveV119898 = 119860V119898119909V119898 +119861V119898119906V119898
119910V119898 = 119862V119898119909V119898
119909V119898 = [119903 120573]
(16)
And 119906V119898 is considered the input to the linear vehicle modelwhile the yaw rate (119903) and vehicle body slip angle (120573) arethe output of the linear vehicle model The parameter stateis written as follows119860V119898
=
[[[[[[
[
(
(1198712119891119862119891+ 119871
2119903119862119903)
119868119904119881
) minus
(119871119903119862119903minus 119871119891119862119891)
119868119904
1 + (
(119871119903119862119903minus 119871119891119862119891)
1198981198812 ) minus
(119862119891+ 119862119903)
119898119881
]]]]]]
]
119861V119898 =[[[
[
119871119891119862119891
119868119904
119862119891
119898119881
]]]
]
119906V119898 = 120575119891
(17)
The parameter of a linear vehicle model directly can bemeasured but some parameters need to be estimated Thereare several methods to estimate themodel parameter [22 23]A method based on Sahoo et al [22] is used to estimate themodel parameters with assumption that the vehicle mass isdistributed to front and rear of each tire The total of vehiclemass (119898) that is based on the front (119898
119891) and rear (119898
119903) tire
masses is written as follows
119898 = 119898119891+119898119903
(18)
when
119898119891= 119898119891119897+119898119891119903
119898119903= 119898119903119897+119898119903119903
(19)
where 119898119891119897 119898119891119903 119898119903119897 and 119898
119903119903are the masses of the vehicle
at front left front right rear left and rear right respectivelyThe parameter length of wheel base (119897) which is sum of lengthfrom a center of front tire 119897
119891and rear tire 119897
119903to COG of vehicle
body is written as follows
119897 = 119897119891+ 119897119903 (20)
while the inertia of vehicle 119868119904that is approximated by assum-
ing the vehicle as two point masses joined with mass-lessroad [22] is written as follows
119868119904= 1198981198911198972119891+1198981199031198972119903 (21)
The front cornering stiffness 119862119891and rear cornering stiffness
119862119903parameters are written in (22) by assuming that the type
of radial tire is used and cornering stiffness per degree of slipangle is approximately 16-17 of the load on the tire [22]Consider
119862119891= 119898119891lowast 119892 lowast 0165 (Ndeg)
119862119903= 119898119903lowast 119892 lowast 0165 (Ndeg)
(22)
The parameters of the linear vehicle model are illustrated inTable 3
3 Force Feedback Torque EstimationControl Algorithm
It has been known that the force feedback torque for driversteering feel should be created artificially due to the elimina-tion of mechanical column shaft in SBW system
Moreover the proportion of the steering feel obtainedshould be similar to conventional steering system For thisreason the steering wheel in SBW system that is equippedwith the motor actuator is used to generate the steeringfeel by controlling the total feedback torque (120591total feedback)A schematic block diagram of the proposed force feedbacktorque control is shown in Figure 5
Based on Figure 5 the total force feedback torque consistsof the torque of front axle motor (120591
119891119898) torque on steering
wheel motor (120591119904119898) and compensation torque which is the
inertia (120591inertia) and damping torque (120591damp) Meanwhile the
6 International Journal of Vehicular Technology
+
+
+
minus
minus
Wheel synchronization
120575sw
120575sw
120591a
Frontaxle
system
LQRcontroller
LQRcontroller
120575m2
120575m2
120575f
ia2
120573
r
Linearvehiclemodel
Force feedback torque
120591map
120591fm
120591sm
120591damp
120591inertia
VGain
scheduling
kfm
kfeelsum
120591total feedback Steeringwheelsystem
Steering wheel system
Steeringwheel
120575m1
120575m1
ia1
ksm
Estimation selfaligning torque
Torque(RMS)
Figure 5 Overview of force feedback torque control algorithm for SBW system
Table 3 Linear vehicle model [16]
Parameter Description Value Unit119862119891 Front cornering stiffness 40000 Nrad
119862119903 Rear cornering stiffness 35000 Nrad
119897119891
Length from center frontwheel to (COG) vehicle 14 m
119897119903
Length from center rearwheel to COG vehicle 1 m
119898 Mass of the vehicle 1535 kg119868119904 Inertia of vehicle 2149 kgm2
119881 Speed of vehicle mdash kmh119865119910119891 Front tire lateral force mdash Nm
119865119910119903 Rear tire lateral force mdash Nm
120572119891 Front tire slip angle mdash Nm
120572119903 Rear tire slip angle mdash Nm
torque map is based on the function of the steering wheelangle and vehicle speed is added to the total feedback torqueMoreover a steering feel gain (119896feel) is adapted based on thefunction of steering wheel angle and vehicle speed to varythe feedback torque In order to control the feedback torquethe gain scheduling with LQR controller is used To showthe influence of the tire on the road surface the self aligningtorque (120591
119886) acts as input disturbance to the front axle system
as follows
120591119886= minus119862
119891(119905119901+ 119905119898) [120573+
119897119891119903
119881minus 120575119891]120583 (23)
where 120583 is coefficient for road dry condition pneumatic trail(119905119901) and mechanical trail (119905
119898)
On the other hand any change of the road surface couldchange the feedback torque magnitude of the front axlemotor Furthermore the LQR controller of the front axle sys-tem is used to synchronize the front tire angle (120575
119891) with driver
input steering wheel angle (120575119904119908) by controlling the angle of
the front axle motor (1205751198982) in accordance with the steering
ratioThus based on the proposed control algorithm the total
force feedback torque to the steering wheel system is writtenas follows
120591total feedback = ([(radicmean (120591119904119898
+ 120591119891119898
+ 120591map))]
+ 120591inertia + 120591damp) 119896feel
(24)
Author [9 10] has generated and discussed the force feed-back torque using the torque map conceptThe torque map isbased on the function of the steering wheel angle and vehiclespeed parameter The feedback torque is determined by twocontrol parameters (119896
120573) for the vehicle speed and (119896
120572) for the
steering wheel angle functionThe torque based on the function of vehicle speed is given
as follows [9]
1205911 = 1198961205731199092(13119909minus
12119881max)+119879in (25)
International Journal of Vehicular Technology 7
0 2 4 6 8 10minus40
minus30
minus20
minus10
0
10
20
30
40
Time (s)
Stee
ring
torq
ue (N
m)
70kmh75kmh
50kmh40kmh30kmh20kmh10kmh
80kmh
60kmh
Figure 6 Steering torque based on torque map at different speed
where (119909) is initial velocity and (119881max) is maximum velocitywhile the torque based on a function of steering wheel angleis written as follows [9]
1205912 = 119896120572radic1003816100381610038161003816120575119904119908
1003816100381610038161003816 sgn 120575119904119908 (26)
Based on the torque map concept the results of steeringtorque at different vehicle speed are shown in Figure 6 whenthe input steering wheel angle is a double lane changemanoeuvreThe torque increased at high speed anddecreasedat low speed [9 24] By varying the two control parametersthe steering reactive torque changes at different torque mag-nitude However the feedback torque based on the torquemap concept did not create the artificial steering feel dueto the fact that no element of road contact such as vehicledynamic responses are feedback to the torque map modelThus driver not sense realistic driver steering feel
Thus for this reason the front axle motor torque andsteering wheel motor torque are added to the torque mapThese ensure that the driver senses the feedback torque fromroad contact The front axle motor torque is chosen due toits proportion to the road surface The reaction torque of theroad surface gives the effect to the rack and pinion systemand it connects to the front axle DC motor Thus the frontaxle motor torque would always be proportional to the roadsurface Any changes in road surface will influence the frontaxle motor torque In order to measure the torque at the frontaxle motor the direct currentmeasurement technique is usedto estimate the torque
The direct relationship between torque and current of amotor can be defined [25]Thus then front axlemotor torque(120591119891119898
) that consists of the front axle motor constant (119896119891119898
) andthe motor current (119894
1198862) is written as follows
120591119891119898
= 1198941198862119896119891119898 (27)
and the corresponding steeringwheelmotor torque (120591119904119898) that
consists of the steering wheel motor constant (119896119904119898) and the
motor current (1198941198861) is written as follows
120591119904119898
= 1198941198861119896119904119898 (28)
0 2 4 6 8 10Time (s)
Mot
or to
rque
(Nm
)
Front axle motor torque
Steering wheel motor torque6
4
2
0
minus2
minus4
minus6
Figure 7 Steering wheel and front axle motor torque at 80 kmh
Figure 7 shows the front axle motor and steering wheelmotor torque response when the input is a double lanechangemanoeuvreThe results show that the front axlemotorgives a high torque This is due to the fact that the inputdisturbance torque between the tire and road surface which isthe self aligning torque has an effect on the front axle motorIncreasing the vehicle speedwill increase the front axlemotortorque On the other hand the magnitude of steering wheelmotor torque is decreased due to the elimination of mechan-ical column shaft and is not affected by the self aligningtorque
The phase compensation torque consists of the inertiatorque and damping torque potential to recreate the artificialsteering feel of the conventional steering system Moreoverit is intended to adjust the steering feel reduce the vibrationand stabilize the system Due to this fact the element of com-pensation torque is taken in order to improve the feedbacktorque for a driver steering feel and to stabilize the system
The damping torque (120591damp) which is the rate change ofthe steering wheel motor angle ( 120575
1198981) and the damping gain(119896damp) is written as follows
120591damp = 1205751198981119896damp (29)
The inertia torque (120591inertia) that is the acceleration of thesteering wheel motor angle ( 120575
1198981) and the inertia gain (119896inertia)is written as follows
120591inertia = 1205751198981119896inertia (30)
The inertia and damping torque responses are shown inFigure 8 when the input steering wheel angle is a double lanechange manoeuvre It is obtained that the damping torquegives a smooth signal and reduces the vibration Meanwhilethe inertia torque gives high magnitude Thus the combina-tion between the inertia torque and damping torque not onlyprovides an average of steering feel but is also able to stabilizethe system at the same time Figure 9 shows the feedback
8 International Journal of Vehicular Technology
0 2 4 6 8 10minus6
minus4
minus2
0
2
4
6
Time (s)
Torq
ue (N
m)
Inertia torque Damping torque
Figure 8 Inertia and damping torque
0 2 4 6 8 10Time (s)
Forc
e fee
dbac
k to
rque
(Nm
)
15
10
5
0
minus5
minus10
minus15
minus20
minus25
10kmh
50kmh100 kmh120 kmh
135kmh
Figure 9 Proposed force feedback torque response at differentvehicle speed
torque response based on (24) The torque is increasedaccording to the vehicle speed The higher vehicle speed willincrease the feedback torque and vice versa due to the samefact that is proposed by torque map
31 Linear Quadratic Regulator (LQR) Controller The LQRcontroller provides the best possible performance to controland stabilize the system by changing the location of the polesof the system to the optimal location for a time responseovershoot and steady stateThis is done by designing the statefeedback control119870 thatminimized cost function 119869 to stabilizethe system In this paper the LQR controller is used firstlyto control the front axle motor angle for the wheel synchro-nization and secondly to control the steering wheel motorfor a feedback torque
The wheel synchronization or directional control is abasic characteristic for steering function where a front tire
+
minus
sum
k1
k2
k3
x1
x2
x3
Plantr(t)
LQR controller
Figure 10 Basic control block diagram of LQR controller
angle should follow driver input steering wheel angle accord-ing to the steering ratio This is done by controlling theangle of the front axle motor In the LQR control schemesa feedback gain matrix is designed to achieve some compro-mise between the use of control effort the magnitude andthe speed of response that will guarantee a stable systemFigure 10 shows the control block diagram of the LQRcontroller
The optimal LQRproblem is often definedmore generallyand consists of finding controller transfer matrix that mini-mizes the cost function 119869 that is written as follows
119869 = int
prop
0(119909119879
119876119909+119906119879
119877119906) 119889119905 (31)
where 119909119879 and119880119879 respectively represent the transpose of the
state and input vector The 119876 and 119877 are positive definite (orpositive-semi definite) real symmetric matrix The 119876 and 119877
are usually chosen to be diagonal matrix written as follows
119876119894119894=
1maximum acceptable value of 1199092
119894
119877119895119895
=1
maximum acceptable value of 1199062119895
(32)
where
119894 = 120576 1 2 ℓ
119895 = 120576 1 2 119895 (33)
In essence Brysonrsquos rule scales the variables that appear in 119869 sothat the maximum acceptable value for each term is oneThisis especially important when the units used for the differentcomponents of 119906 and 119909 make the values for these variablesnumerically very different from each other
Assume that the steering ratio is 15 1 and the gains ofthe LQR controller are [01 098 0008] Figure 11 shows aresponse of the front axle system when the input steeringwheel angle is a step function as shown in Figure 11(a)
The results show that the angle of the front axle motorhas a fast response compared to the EPS steering system
International Journal of Vehicular Technology 9
0 2 4 6 8 100
2
4
6
8
10
12
14
16
18
20
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle
0 2 4 6 8 100
2
4
6
8
10
12
14
16
18
Time (s)
Mot
orc
olum
n an
gle (
deg)
EPS steeringmdashcolumn shaft angle
SBW systemmdashmotor angle
(b) Front axle motor angle
0 2 4 6 8 10
Rack
disp
lace
men
t (m
)
SBW system
EPS steering
Time (s)
18
16
14
12
10
8
6
4
2
0
times10minus3
(c) Rack displacement
0 2 4 6 8 100
02
04
06
08
1
12
14Fr
ont t
ire an
gle (
deg)
SBW system
EPS steering
Time (s)
(d) Front tire angle
Figure 11 Front axle system response for wheel synchronization
Moreover the LQR controller is able to reduce the delaycaused by the self aligning torque created from the contactbetween the tire and road surfaces This not only improvedthe vehicle stability but also ensured that the driver has amore confident level during manoeuvre Compared to theEPS steering system the delay caused by mechanical columnshaft connected to the steering wheel system Furthermore itcan be seen that the front tire angle is at almost fairly close to1∘ as shown in Figure 11(d) However it has small chatteringdue to the self aligning torque However with wide rangeadjustment gains of the LQR controller the steering responsecould be improved
32 Gain Scheduling (GS) with LQR Controller In certainsituations the dynamic behaviour of the process changed inaccordance with the conditions of the process This situation
could affect the stability of the system if not to control prop-erly Thus it is possible to change the gains of the controllerby monitoring the operating conditions of the process tohave a stable system The gain scheduling (GS) techniqueis used to change the gain of the LQR controller in order tocontrol the feedback torque and the steering feel gain (119896feel)is used to increase and decrease the torque It is important toincrease the feedback torque at high speed for steering stabil-ity and to decrease it in low speed especially during parkingto improve vehicle manoeuvre [9] This is based on the steer-ingwheel angle and vehicle speed function in order to achievethe feedback torque for the real driver steering feel
The relationship between a gain scheduling with LQRcontroller is shown in Table 4 The vehicle speed is dividedinto 3 categories at low speed (0ndash30 kmh) medium speed(31ndash100 kmh) and high speed (101ndash120 kmh) The maxi-mum vehicle speed is limited to 120 kmhThe steering wheel
10 International Journal of Vehicular Technology
0 2 4 6 8 10minus50
minus40
minus30
minus20
minus10
0
10
20
30
40
50
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Input steering wheel angle
0 1 2 3 4 5 6 7 8 9 10minus15
minus10
minus5
0
5
10
Time (s)
Stee
ring
torq
ue (N
m) LQR
LQR + GS
Reference
(b) 120583 = 1
0 1 2 3 4 5 6 7 8 9 10
minus15
minus10
minus5
0
5
10
15
Stee
ring
torq
ue (N
m) LQR
Reference
Time (s)
LQR + GS
(c) 120583 = 14
0 1 2 3 4 5 6 7 8 9 10minus20
minus15
minus10
minus5
0
5
10
15
20
Stee
ring
torq
ue (N
m) LQR LQR + GS
Reference
Time (s)
(d) 120583 = 17
Figure 12 Comparison of torque response of LQR and LQR + GS controller at different 120583
Table 4 The gains of LQR controller and steering feel
119881 (kmh) 120575119904119908
(degree) 1198961 1198962 1198963 119896feel
(0ndash30) (plusmn180∘ lt 120575119904119908
lt plusmn405∘) 003 065 0001 014(31ndash100) (plusmn46∘ lt 120575
119904119908lt plusmn179∘) 003 084 0001 060
(101ndash120) 120575119904119908
le +45∘ or 120575119904119908
le minus45∘ 005 095 0001 095
angle is also divided into 3 categories according to vehiclespeed condition The notation ldquo+rdquo means that the driverturned the steering wheel clockwise and notation ldquominusrdquo is usedfor counter clockwise in which the maximum steering wheelturn is plusmn405∘
The gains of the LQR controller are defined using Brysonrsquosrule method and the steering feel gain is defined usingintuitive and has a range from 01 to 1 This range has beendefined for various speeds 10 kmh 80 kmh and 120 kmhafter comparing with EPS steering system [15] For exampleif the driver turns a steering wheel between (plusmn180∘ lt 120575
119904119908lt
plusmn405∘) and vehicle speed is 10 kmh the total feedback torque
is multiplied by steering feel gain 119896feel = 014 Thus thefeedback torque is decreased which results in making thedriver feel it easier turning the steering wheel and improvesvehicle manoeuvre The feedback torque is regulated by LQRcontroller when the gains are 1198961 = 003 1198962 = 065 and1198963 = 0001 The gains are the optimum that has been assuredby providing a better torque control based on vehicle speedand steering wheel angle function
Comparing torque response between LQR and LQR +GScontroller for different coefficient (120583) is shown in Figure 12when the input steering wheel angle is a double lane changemanoeuvre as shown in Figure 12(a) The vehicle speed isset to be 80 kmh Based on results obtained the LQR + GStechnique offers an improvement 90 and the curve trend isalmost similar to reference torque compared to the LQR con-troller This is because the gains of LQR controller changedbased on the steering wheel angle and vehicle speed param-eter which is able to reject the uncertainty torque of a roadcondition Meanwhile with the fixed gains of the LQR con-troller it has an excessive torque that can cause the system tobe unstable The RMS torque values are illustrated in Table 5
International Journal of Vehicular Technology 11
Table 5 Comparison of RMS torque values of LQR and LQR + GScontroller
Reference LQR LQR + GS120583 = 1 56361 50795 55450120583 = 14 61218 68339 57243120583 = 17 66184 72888 62080
minus50 0 50minus15
minus10
minus5
0
5
10
15
Steering angle (deg)
Stee
ring
torq
ue (N
m) Feel gain = 5
Feel gain = 01Feel gain = 05
Feel gain = 1
(a) Different steering feel gain at 80 kmh
minus50 minus40 minus30 minus20 minus10 0 10 20 30 40 50
minus15
minus10
minus5
0
5
10
15
Steering wheel angle (deg)
Stee
ring
torq
ue (N
m)
80kmh kfeel = 060
120 kmh kfeel = 095
10kmh kfeel = 014
(b) Different steering feel gain at various speed
Figure 13 Steering feel gains response at various vehicle speed
Figure 13(a) shows the steering torque with respect tosteering wheel angle when different steering feel gain is usedat various vehicle speeds It is observed that when the steeringfeel gain is increased at maximum range which is 1 thesteering torque will increase and driver will sense stiffer onthe steering wheel at high speed manoeuvre Meanwhilewhen decreasing the steering feel gain this will result inreducing the steering torqueThus the driver will sense softer
Target monitor
Host PC
Cross over cableConnector block
PCI 6221
Power supplyCurrent sensor
DC motor driver
Figure 14 Steering wheel system hardware overview
during turning the steering wheel at the low speed manoeu-vre However when the steering feel gain is 5 which exceedsits maximum limit the steering torque becomes unstableFigure 13(b) shows the steering feel gains at various speedsAt high speed of 120 kmh the steering feel gain is increasedand this will increase the steering torque for steering stabilityMeanwhile at low speed the steering feel gain is decreasedwhich will help improve vehicle manoeuvre especially duringparking Based on the proposed gain scheduling with theLQR controller not only does it provide better control on thesteering torque but also it is optional for a designer to have awide range adjustment to vary the steering torque for a betterdriver steering feel and to improve vehicle manoeuvre
4 Hardware in the Loop (HIL) and Result
For hardware in the loop (HIL) setup the front axle systemand linear vehicle model are simulated on the host computerThe steering wheel system is replaced by HIL mechanismThe position of the steering wheel motor is measured usinga rotary encoder with 500 pulses per revolution To measurethe steering wheel torque the current sensor is connectedin series with the DC motor driver The steering wheel DCmotor has a torque 129Nm with 100 rpm The single tracklinear vehicle model is used to generate the estimation selfaligning torque as element to the feedback torque and distur-bance to the front axle system The HIL system is executedunder 1ms sampling time and interfaced using Matlab-XPCtarget toolbox configuration as shown in Figures 14 and 15
The assumptions through the experiment are listed asfollows
(i) the vehicle cruises at constant speed(ii) the condition of dry asphalt road (120583 = 085)
The experiments are done under high medium and lowspeed manoeuvre [26] The same input steering wheel anglebetween simulation and experiment is given Therefore theresult obtained from real time control is compared with theelectronic power steering (EPS) system [15]
41 Motor Model Validation for Steering Wheel SystemAlthough the model of steering wheel system has been
12 International Journal of Vehicular Technology
Stee
ring
whe
el sy
stem
DC motor
Current
Encoder
16 bit card PCI 6221
Analogoutput
Analoginput
Counterchannel
TargetPC
Monitortarget
PC
Crossover cable
Host PC
Front axle systemand
linear vehicle model
Hardware part Simulation part
Figure 15 Block diagram of the steering wheel-XPC target interfaced
Steering wheel model
120575m1
120575m1
ia1
r(t)
Figure 16 Block diagrammotormodel validation for steeringwheelsystem
derived the model validation is necessary to ensure thesystem is fairly able to adequate for real time applicationThis section describes the motor model validation of steeringwheel system The block diagram is shown in Figure 16
The input signal is 119903(119905) and the output is rate change ofmotor ( 120575
1198981) motor position (1205751198981) and the motor current
( 1198941198981) The measured output response is then used and com-
pared with simulation to estimate the parameter of the steer-ing wheel model using the Matlab tool software The mea-sured output data are taken based on the open loop systemwhen the input to the system is sinusoidal signal withamplitude of plusmn45∘ for 10 sec The Gaussian method is usedto estimate the parameters of the steering wheel motorThe comparison between simulation and measured outputis shown in Figure 17 It is obtained that the output of thesimulation is almost identical to the measured output How-ever small difference errors occur due to the small parametricuncertainties that have not been modelled The parameterestimation value is illustrated in Table 6
42 Noise Filtering Both analog and digital devices have atrait of noise or unwanted features that have effect on thesystem performance such as electromagnetic interferenceThe noise could be either random or white noise Attenuationof this noise is often as a primary goal in control systemapplication and in particular when controlling the processsystem The average value of noise typically is zero whichwill give misreading information to the process and it wouldnot be possible to control the process In order to reduce thenoise inference filtering technique is used A filter is used toremove unwanted noise It removes some noise frequenciesto suppress interfacing signal and reduce background noise
Table 6 Parameter estimation for DCmotor steering wheel system
Parameter Value Estimation value Unit1198771 4405 564 ohm1198711 0010 0017 henry119870119904119898
1459 0024 Nm1198691198981 0002 00036 (kgm2)1198871198981 0004 00068 Nm(rads)1198961199041 165 0025 Nm
There are many types of filtering techniques such as low passfilter that has been used in this study The low pass filter isused to filter out the interference noise from the output of thecurrent sensor
Figure 18 shows the comparison of the output responsefor a current sensor with and without using filters Withouta filter the current sensor has high amplitude interferencenoise However with low pass filter under the frequency of5Hz the interference noise is reduced and the fluctuations areremoved Thus this provides a smoother output response
43 Effect on Compensation Torque A steering torque com-parison without and with compensation torque is shown inFigure 19The results for without compensation torque showsthat the steering torque response curve is far away fromEPS steering system with different amplitude as shown inFigure 19(a) This happens due to less of derivative reactiontorque from the steering wheel motor Figure 19(b) showssteering torque response with compensation torque
It is observed that by adding the additional compensationtorque the steering torque response curve trends fairly closeto EPS steering system Thus it helps to improve the driversteering feel during manoeuvre and stabilize the systemHowever with the wide range adjustment gains of the com-pensation torque it could improve the steering torqueresponse
44 Medium Speed Manoeuvre The experimental steeringtorque results at medium speed manoeuvre are shown inFigure 20(b) when the steering wheel angle is plusmn100∘ as shown
International Journal of Vehicular Technology 13
0 2 4 6 8 10minus50
0
50
Time (s)
Rate
chan
ge m
otor
pos
ition
(deg
s)
MeasurementSimulation
(a) Rate change of motor position
0 2 4 6 8 100
10
20
30
40
50
60
Time (s)
Mot
or p
ositi
on (d
eg)
Measurement
Simulation
(b) Position of the motor
0 2 4 6 8 10minus01
minus005
0
005
01
015
Mot
or cu
rren
t (A
)
MeasurementSimulation
Time (s)
(c) Current of the motor
Figure 17 Motor model validation comparison between simulation and measurement
With filter Without filter
Time (s)
Curr
ent (
A)
008
006
004
002
0
minus002
minus004
minus006
minus0081 15 2 25 3 35 4 45 5 55
Figure 18 Noise filtering using low pass filter
in Figure 20(a) From the results the experimental steeringtorque response is almost identical to the EPS steering systemHowever it has small different torque amplitudes These
occur due to the fact that mechanism or friction from themechanical gearbox motor is not modelled in proposedcontrol algorithm
14 International Journal of Vehicular Technology
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
minus152 3 4 5 6 7 8 9 10
(a) Without compensation torque
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
21 3 4 5 6 7 8 9 10
(b) With compensation torque
Figure 19 Steering torque influence on compensation torque
minus80
minus60
minus40
minus20
0
20
40
60
80
100
Stee
ring
whe
el an
gle (
deg)
Time (s)3 4 5 6 7 8 9 10
(a) Steering wheel angle at 80 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
15
20
5
0
minus5
minus10
minus15
minus203 4 5 6 7 8 9 10
(b) Steering torque estimation at 80 kmh
Figure 20 Steering torque estimation and steering wheel angle response at 80 kmh
45 High Speed Manoeuvre At high speed manoeuvre theinput steering wheel angle that is more than plusmn45∘ is very dan-gerous [12] Therefore in this experiment the input steeringwheel angle is less than plusmn45∘ as shown in Figure 21(a) Basedon the results shown in Figures 21(b) and 21(c) the experi-mental steering torque is obtained fairly similar to the EPSsteering system Although the steering wheel angle is at smallangles the influence of vehicle speed cause the steering torqueis increased Moreover at higher speed the vehicle has moreunder steer propensity [9] This increased the self aligningtorque in proportional high torque at front axle motorThusthe driver will sense stiffer on the steering wheel duringmanoeuvre
In general the centering of the steering wheel perfor-mance or steering wheel returnability is better when thesteering torque gradient is large and the hysteresis is small [9]Based on this fact the result is shown in Figure 21(c) whereby
the hysteresis of the steering wheel returnability based onexperimental result is smaller compared to EPS steeringsystem In addition the steering torque gradient is larger thanEPS steering torque Therefore it is considered that the pro-posed feedback torque algorithmnot only achieved a steeringfeel but also improved the steering wheel returnability
46 Lower Speed Manoeuvre The steering torque results atlow speed manoeuvre are shown in Figures 22(b) and 22(c)when the input steering wheel is as shown in Figure 22(a)The vehicle cruised at constant speed of 10 kmh
The experimental steering torque is almost comparableto EPS torque but the amplitude is decreases The torque isdecreased due to the fact that the vehicle speed is reducedwhich is effect on reducing the demand to the total feedbacktorque Moreover the large turning of the steering wheelincreases the backlash of the motor which then contributes
International Journal of Vehicular Technology 15
2 3 4 5 6 7 8 9 10minus30
minus20
minus10
0
10
20
30
40
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle at 120 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steeringSimulationExperiment
10
5
0
minus5
minus10
3 4 5 6 7 8 9
(b) Steering torque estimation at 120 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment10
15
5
0
minus5
minus10
minus15minus20 minus10 0 10 20 30 40
Steering wheel angle (deg)
(c) Steering torque estimation against steering wheel angle at 120 kmh
Figure 21 Steering torque estimation response at 120 kmh
to the decrease of the motor torque Thus driver will sensesofter during steering the wheel during maneuver Howeverit has small different torque against EPS steering Wide rangeadjustment of gains in proposed control algorithm may helpimprove the steering torque response
3 4 5 6 7 8 9 10minus150
minus100
minus50
0
50
100
150
200
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
10
15
5
0
minus5
minus10
minus15
Time (s)2 3 4 5 6 7 8 9 10
(b) Steering torque estimation at 10 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
Steering wheel angle (deg)
10
8
6
4
2
0
minus2
minus4
minus6
minus8
minus10
minus100 minus50 0 50 100 150
(c) Steering torque estimation against steering wheel angle at 10 kmh
Figure 22 Steering torque estimation response at 10 kmh
5 Conclusion and Future Work
The proposed force feedback torque estimation and controlalgorithm have been described The proposed control algo-rithm is considered as an effort to generate the feedback
16 International Journal of Vehicular Technology
torque for driver steering feel in the vehicle SBW systemThe mathematical model of SBW subsystem consists ofsteering wheel and front axle systems that are modelledThe effectiveness of the control algorithm is verified throughthe experimental hardware in the loop (HIL) interfacedwith Matlab XPC target toolbox software According to theexperimental results the proposed control algorithm is ableto generate the feedback torque and the driver is able to sensesteering feel similar to that obtained in EPS steering systemThe LQR controller with gain scheduling based on steeringwheel angle and vehicle speed function provides better torquecontrol that allows rejection of the uncertainty torque froma road condition Furthermore the steering feel gain is ableto increase and lower the steering torque with improvedsteering stability at high speed and it improved the vehiclemanoeuvre at low speed It is also found that compensationtorque is able to improve the feedback torque and to stabilizethe system However it has small different torque magnitudethat occurs due to the backlash of steering wheel motor anduncertainty disturbance of mechanical properties Howeverwith a wide range adjustment of the gains on the proposedcontrol algorithm it may improve and provide realisticsteering torque for better driver steering feel
In future work the proposed control algorithm will betested in real vehicle in order to have more realistic steeringtorque for further investigation and analysis
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
Acknowledgments
This work is fully supported by the Ministry of Higher Edu-cation Malaysia and University Technology Malaysia underFRGS (vote 4F370) and Research University Grant (vote00G64) This work is also supported by Proton HoldingBerhad
References
[1] Y Yao ldquoVehicle steer-by-wire system controlrdquo SAE TechnicalPaper 2006-01-1175 2006
[2] E A Kumar D Dinesh and N Kamble ldquoAn overview of activefront steering systemrdquo International Journal of Scientific ampEngineering Research vol 3 no 6 2012
[3] S M H Fahami H Zamzuri S A Mazlan and N B Zulkar-nain ldquoThe design of vehicle active front steering based on steerby wire systemrdquo Advanced Science Letters vol 19 no 1 pp 61ndash65 2013
[4] S M H Fahami H Zamzuri S A Mazlan and S A SaruchildquoThe variable steering ratio for vehicle steer by wire systemusing hyperbolic tangent methodrdquo Applied Mechanics andMaterials vol 575 pp 781ndash784 2014
[5] T Kaufmann S Milsap B Murray and J Petrowski ldquoDevelop-ment experience with steer by wirerdquo in Proceedings of the SAEInternational Congress and Exhibition SAE 2001-01-2479 2001
[6] S Amberkare F Bolourchi J Demerly and S Millsap ldquoAcontrol system methodology for steer by wire systemsrdquo SAETechnical Paper 2004-01-1106 2004
[7] A Baviskar J R Wagner D M Dawson D Braganza andP Setlur ldquoAn adjustable steer-by-wire haptic-interface trackingcontroller for ground vehiclesrdquo IEEE Transactions on VehicularTechnology vol 58 no 2 pp 546ndash554 2009
[8] M Segawa S Kimura T Kada and S Nakona ldquoA study of reac-tive torque control for steer bywire systemrdquo inProceedings of theInternational Symposium on Advanced Vehicle Control (AVECrsquo02) pp 653ndash658 Hiroshima Japan September 2002
[9] S W Oh H C Chae S C Yun and C S Han ldquoThe design ofa controller for the steer-by-wire systemrdquo JSME InternationalJournal Series C Mechanical Systems Machine Elements andManufacturing vol 47 no 3 pp 896ndash907 2004
[10] C J Kim J H Jang S K Oh J Y Lee and J K HedrickldquoDevelopment of a control algorithm for a rack-actuating steer-by-wire system using road information feedbackrdquo Proceedingsof the Institution of Mechanical Engineers Part D Journal ofAutomobile Engineering vol 222 no 9 pp 1559ndash1571 2008
[11] S Asai H Kuroyanagi and S Takeuchi ldquoDevelopment of asteer-by-wire system with force feedback using a disturbanceobserverrdquo inProceedings of the SAEWorldCongressampExhibitionon Steering amp Suspension Technology Detroit Mich USA 2004
[12] T J Park C S Han and S H Lee ldquoDevelopment of the elec-tronic control unit for the rack-actuating steer-by-wire usingthe hardware-in-the-loop simulation systemrdquo Mechatronicsvol 15 no 8 pp 899ndash918 2005
[13] D Odenhtal T Bunte H D Heitzer and C Eicker ldquoHow tomake steer-by-wire feel like power steeringrdquo in Proceedings ofthe 15th IFACWorld Congress Barcelona Spain July 2000
[14] H Na C Zong and D Hu ldquoInvestigations on corneringcontrol algorithm design and road feeling optimization for asteer-by-wire vehiclerdquo in Proceedings of the IEEE InternationalConference on Mechatronics and Automation (ICMA rsquo09) pp3246ndash3251 Changchun China August 2009
[15] S Ancha A Baviskar J RWagner andDMDawson ldquoGroundvehicle steering systems modelling control and analysis ofhydraulic electric and steer-by-wire configurationsrdquo Interna-tional Journal of Vehicle Design vol 44 no 1-2 pp 188ndash2082007
[16] P Yih and J C Gerdes ldquoModification of vehicle handlingcharacteristics via steer-by-wirerdquo IEEE Transactions on ControlSystems Technology vol 13 no 6 pp 965ndash976 2005
[17] S M H Fahami H Zamzuri S A Mazlan and M A ZakarialdquoModeling and simulation of vehicle steer by wire systemrdquoin Proceedings of the IEEE Symposium on Humanities Scienceand Engineering Research (SHUSER rsquo12) pp 765ndash770 KualaLumpur Malaysia June 2012
[18] S Anwar and L Chen ldquoAn analytical redundancy-based faultdetection and isolation algorithm for a road-wheel controlsubsystem in a steer-by-wire systemrdquo IEEE Transactions onVehicular Technology vol 56 no 5 pp 2859ndash2869 2007
[19] M T Do Z Man C Zhang H Wang and F S Tay ldquoRobustsliding mode-based learning control for steer-by-wire systemsin modern vehiclesrdquo IEEE Transactions on Vehicular Technol-ogy vol 63 no 2 pp 580ndash590 2014
[20] J Duan R Wang and Y Yu ldquoResearch on control strategies ofsteer-by-wire systemrdquo in Proceedings of the International Con-ference on Intelligent Computation Technology and Automation(ICICTA rsquo10) vol 2 pp 1122ndash1125 May 2010
International Journal of Vehicular Technology 17
[21] B Zheng and S Anwar ldquoYaw stability control of a steer-by-wireequipped vehicle via active front wheel steeringrdquoMechatronicsvol 19 no 6 pp 799ndash804 2009
[22] S Sahoo S C Subramanian and S Srivastava ldquoDesign andimplementation of a controller for navigating an autonomousground vehiclerdquo in Proceedings of the 2nd International Confer-ence on Power Control and Embedded Systems (ICPCES rsquo12) pp1ndash6 IEEE Allahabad India December 2012
[23] T D Gillespie Fundamentals of Vehicle Dynamics Society ofAutomotive Engineers 1992
[24] M H Lee S K Ha J Y Choi and K S Yoon ldquoImprovement ofthe steering feel of an electric power steering system by torquemap modificationrdquo Journal of Mechanical Science and Technol-ogy vol 19 no 3 pp 792ndash801 2005
[25] R Pastorino M A Naya J A Perez and J Cuadrado ldquoGearedPM coreless motor modelling for driverrsquos force feedback insteer-by-wire systemsrdquo Mechatronics vol 21 no 6 pp 1043ndash1054 2011
[26] P Koehn andM Eckrich ldquoActive steeringmdashthe BMWapproachtowards modern steering technologyrdquo SAE Technical PaperSeries 2004-01-1105 SAE International 2004
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
International Journal of Vehicular Technology 5
With the aforementioned assumptions the derived equationsof the motion of yaw rate (119903) and vehicle body slip angle (120573)are shown in (13) as follows
119903 = [
[
(1198972119891119862119891+ 119897
2119903119862119903)
119868119904119881
]
]
119903+[
[
(1198972119903119862119903minus 119897
2119891119862119891)
119868119904
]
]
120573
+[
(119897119891119862119891)
119868119904
]120575119891
120573 = [1+(119897119903119862119903minus 119897119891119862119891)
1198981198812 ] 119903 minus[
(119862119891+ 119862119903)
119898119881]120573
+[
119862119891
119898119881]120575119891
(13)
In the linear tire region lateral force at the front and rear isrelated to slip angle by cornering stiffness of the front and reartires written as follows
119865119910119891
= minus119862119891120572119891
119865119910119903
= minus119862119903120572119903
(14)
At a larger slip angle the relation between lateral force andslip angle becomes nonlinear By assuming a smaller slipangle approximation the slip angle is written in terms ofvehicle body slip angle and yaw rate that are considefed asfollows
120572119891= (120573+
119897119891
119881119903minus 120575119891)
120572119903= (120573+
119897119903
119881119903minus 120575119903)
(15)
Rewriting (13) into state space form we haveV119898 = 119860V119898119909V119898 +119861V119898119906V119898
119910V119898 = 119862V119898119909V119898
119909V119898 = [119903 120573]
(16)
And 119906V119898 is considered the input to the linear vehicle modelwhile the yaw rate (119903) and vehicle body slip angle (120573) arethe output of the linear vehicle model The parameter stateis written as follows119860V119898
=
[[[[[[
[
(
(1198712119891119862119891+ 119871
2119903119862119903)
119868119904119881
) minus
(119871119903119862119903minus 119871119891119862119891)
119868119904
1 + (
(119871119903119862119903minus 119871119891119862119891)
1198981198812 ) minus
(119862119891+ 119862119903)
119898119881
]]]]]]
]
119861V119898 =[[[
[
119871119891119862119891
119868119904
119862119891
119898119881
]]]
]
119906V119898 = 120575119891
(17)
The parameter of a linear vehicle model directly can bemeasured but some parameters need to be estimated Thereare several methods to estimate themodel parameter [22 23]A method based on Sahoo et al [22] is used to estimate themodel parameters with assumption that the vehicle mass isdistributed to front and rear of each tire The total of vehiclemass (119898) that is based on the front (119898
119891) and rear (119898
119903) tire
masses is written as follows
119898 = 119898119891+119898119903
(18)
when
119898119891= 119898119891119897+119898119891119903
119898119903= 119898119903119897+119898119903119903
(19)
where 119898119891119897 119898119891119903 119898119903119897 and 119898
119903119903are the masses of the vehicle
at front left front right rear left and rear right respectivelyThe parameter length of wheel base (119897) which is sum of lengthfrom a center of front tire 119897
119891and rear tire 119897
119903to COG of vehicle
body is written as follows
119897 = 119897119891+ 119897119903 (20)
while the inertia of vehicle 119868119904that is approximated by assum-
ing the vehicle as two point masses joined with mass-lessroad [22] is written as follows
119868119904= 1198981198911198972119891+1198981199031198972119903 (21)
The front cornering stiffness 119862119891and rear cornering stiffness
119862119903parameters are written in (22) by assuming that the type
of radial tire is used and cornering stiffness per degree of slipangle is approximately 16-17 of the load on the tire [22]Consider
119862119891= 119898119891lowast 119892 lowast 0165 (Ndeg)
119862119903= 119898119903lowast 119892 lowast 0165 (Ndeg)
(22)
The parameters of the linear vehicle model are illustrated inTable 3
3 Force Feedback Torque EstimationControl Algorithm
It has been known that the force feedback torque for driversteering feel should be created artificially due to the elimina-tion of mechanical column shaft in SBW system
Moreover the proportion of the steering feel obtainedshould be similar to conventional steering system For thisreason the steering wheel in SBW system that is equippedwith the motor actuator is used to generate the steeringfeel by controlling the total feedback torque (120591total feedback)A schematic block diagram of the proposed force feedbacktorque control is shown in Figure 5
Based on Figure 5 the total force feedback torque consistsof the torque of front axle motor (120591
119891119898) torque on steering
wheel motor (120591119904119898) and compensation torque which is the
inertia (120591inertia) and damping torque (120591damp) Meanwhile the
6 International Journal of Vehicular Technology
+
+
+
minus
minus
Wheel synchronization
120575sw
120575sw
120591a
Frontaxle
system
LQRcontroller
LQRcontroller
120575m2
120575m2
120575f
ia2
120573
r
Linearvehiclemodel
Force feedback torque
120591map
120591fm
120591sm
120591damp
120591inertia
VGain
scheduling
kfm
kfeelsum
120591total feedback Steeringwheelsystem
Steering wheel system
Steeringwheel
120575m1
120575m1
ia1
ksm
Estimation selfaligning torque
Torque(RMS)
Figure 5 Overview of force feedback torque control algorithm for SBW system
Table 3 Linear vehicle model [16]
Parameter Description Value Unit119862119891 Front cornering stiffness 40000 Nrad
119862119903 Rear cornering stiffness 35000 Nrad
119897119891
Length from center frontwheel to (COG) vehicle 14 m
119897119903
Length from center rearwheel to COG vehicle 1 m
119898 Mass of the vehicle 1535 kg119868119904 Inertia of vehicle 2149 kgm2
119881 Speed of vehicle mdash kmh119865119910119891 Front tire lateral force mdash Nm
119865119910119903 Rear tire lateral force mdash Nm
120572119891 Front tire slip angle mdash Nm
120572119903 Rear tire slip angle mdash Nm
torque map is based on the function of the steering wheelangle and vehicle speed is added to the total feedback torqueMoreover a steering feel gain (119896feel) is adapted based on thefunction of steering wheel angle and vehicle speed to varythe feedback torque In order to control the feedback torquethe gain scheduling with LQR controller is used To showthe influence of the tire on the road surface the self aligningtorque (120591
119886) acts as input disturbance to the front axle system
as follows
120591119886= minus119862
119891(119905119901+ 119905119898) [120573+
119897119891119903
119881minus 120575119891]120583 (23)
where 120583 is coefficient for road dry condition pneumatic trail(119905119901) and mechanical trail (119905
119898)
On the other hand any change of the road surface couldchange the feedback torque magnitude of the front axlemotor Furthermore the LQR controller of the front axle sys-tem is used to synchronize the front tire angle (120575
119891) with driver
input steering wheel angle (120575119904119908) by controlling the angle of
the front axle motor (1205751198982) in accordance with the steering
ratioThus based on the proposed control algorithm the total
force feedback torque to the steering wheel system is writtenas follows
120591total feedback = ([(radicmean (120591119904119898
+ 120591119891119898
+ 120591map))]
+ 120591inertia + 120591damp) 119896feel
(24)
Author [9 10] has generated and discussed the force feed-back torque using the torque map conceptThe torque map isbased on the function of the steering wheel angle and vehiclespeed parameter The feedback torque is determined by twocontrol parameters (119896
120573) for the vehicle speed and (119896
120572) for the
steering wheel angle functionThe torque based on the function of vehicle speed is given
as follows [9]
1205911 = 1198961205731199092(13119909minus
12119881max)+119879in (25)
International Journal of Vehicular Technology 7
0 2 4 6 8 10minus40
minus30
minus20
minus10
0
10
20
30
40
Time (s)
Stee
ring
torq
ue (N
m)
70kmh75kmh
50kmh40kmh30kmh20kmh10kmh
80kmh
60kmh
Figure 6 Steering torque based on torque map at different speed
where (119909) is initial velocity and (119881max) is maximum velocitywhile the torque based on a function of steering wheel angleis written as follows [9]
1205912 = 119896120572radic1003816100381610038161003816120575119904119908
1003816100381610038161003816 sgn 120575119904119908 (26)
Based on the torque map concept the results of steeringtorque at different vehicle speed are shown in Figure 6 whenthe input steering wheel angle is a double lane changemanoeuvreThe torque increased at high speed anddecreasedat low speed [9 24] By varying the two control parametersthe steering reactive torque changes at different torque mag-nitude However the feedback torque based on the torquemap concept did not create the artificial steering feel dueto the fact that no element of road contact such as vehicledynamic responses are feedback to the torque map modelThus driver not sense realistic driver steering feel
Thus for this reason the front axle motor torque andsteering wheel motor torque are added to the torque mapThese ensure that the driver senses the feedback torque fromroad contact The front axle motor torque is chosen due toits proportion to the road surface The reaction torque of theroad surface gives the effect to the rack and pinion systemand it connects to the front axle DC motor Thus the frontaxle motor torque would always be proportional to the roadsurface Any changes in road surface will influence the frontaxle motor torque In order to measure the torque at the frontaxle motor the direct currentmeasurement technique is usedto estimate the torque
The direct relationship between torque and current of amotor can be defined [25]Thus then front axlemotor torque(120591119891119898
) that consists of the front axle motor constant (119896119891119898
) andthe motor current (119894
1198862) is written as follows
120591119891119898
= 1198941198862119896119891119898 (27)
and the corresponding steeringwheelmotor torque (120591119904119898) that
consists of the steering wheel motor constant (119896119904119898) and the
motor current (1198941198861) is written as follows
120591119904119898
= 1198941198861119896119904119898 (28)
0 2 4 6 8 10Time (s)
Mot
or to
rque
(Nm
)
Front axle motor torque
Steering wheel motor torque6
4
2
0
minus2
minus4
minus6
Figure 7 Steering wheel and front axle motor torque at 80 kmh
Figure 7 shows the front axle motor and steering wheelmotor torque response when the input is a double lanechangemanoeuvreThe results show that the front axlemotorgives a high torque This is due to the fact that the inputdisturbance torque between the tire and road surface which isthe self aligning torque has an effect on the front axle motorIncreasing the vehicle speedwill increase the front axlemotortorque On the other hand the magnitude of steering wheelmotor torque is decreased due to the elimination of mechan-ical column shaft and is not affected by the self aligningtorque
The phase compensation torque consists of the inertiatorque and damping torque potential to recreate the artificialsteering feel of the conventional steering system Moreoverit is intended to adjust the steering feel reduce the vibrationand stabilize the system Due to this fact the element of com-pensation torque is taken in order to improve the feedbacktorque for a driver steering feel and to stabilize the system
The damping torque (120591damp) which is the rate change ofthe steering wheel motor angle ( 120575
1198981) and the damping gain(119896damp) is written as follows
120591damp = 1205751198981119896damp (29)
The inertia torque (120591inertia) that is the acceleration of thesteering wheel motor angle ( 120575
1198981) and the inertia gain (119896inertia)is written as follows
120591inertia = 1205751198981119896inertia (30)
The inertia and damping torque responses are shown inFigure 8 when the input steering wheel angle is a double lanechange manoeuvre It is obtained that the damping torquegives a smooth signal and reduces the vibration Meanwhilethe inertia torque gives high magnitude Thus the combina-tion between the inertia torque and damping torque not onlyprovides an average of steering feel but is also able to stabilizethe system at the same time Figure 9 shows the feedback
8 International Journal of Vehicular Technology
0 2 4 6 8 10minus6
minus4
minus2
0
2
4
6
Time (s)
Torq
ue (N
m)
Inertia torque Damping torque
Figure 8 Inertia and damping torque
0 2 4 6 8 10Time (s)
Forc
e fee
dbac
k to
rque
(Nm
)
15
10
5
0
minus5
minus10
minus15
minus20
minus25
10kmh
50kmh100 kmh120 kmh
135kmh
Figure 9 Proposed force feedback torque response at differentvehicle speed
torque response based on (24) The torque is increasedaccording to the vehicle speed The higher vehicle speed willincrease the feedback torque and vice versa due to the samefact that is proposed by torque map
31 Linear Quadratic Regulator (LQR) Controller The LQRcontroller provides the best possible performance to controland stabilize the system by changing the location of the polesof the system to the optimal location for a time responseovershoot and steady stateThis is done by designing the statefeedback control119870 thatminimized cost function 119869 to stabilizethe system In this paper the LQR controller is used firstlyto control the front axle motor angle for the wheel synchro-nization and secondly to control the steering wheel motorfor a feedback torque
The wheel synchronization or directional control is abasic characteristic for steering function where a front tire
+
minus
sum
k1
k2
k3
x1
x2
x3
Plantr(t)
LQR controller
Figure 10 Basic control block diagram of LQR controller
angle should follow driver input steering wheel angle accord-ing to the steering ratio This is done by controlling theangle of the front axle motor In the LQR control schemesa feedback gain matrix is designed to achieve some compro-mise between the use of control effort the magnitude andthe speed of response that will guarantee a stable systemFigure 10 shows the control block diagram of the LQRcontroller
The optimal LQRproblem is often definedmore generallyand consists of finding controller transfer matrix that mini-mizes the cost function 119869 that is written as follows
119869 = int
prop
0(119909119879
119876119909+119906119879
119877119906) 119889119905 (31)
where 119909119879 and119880119879 respectively represent the transpose of the
state and input vector The 119876 and 119877 are positive definite (orpositive-semi definite) real symmetric matrix The 119876 and 119877
are usually chosen to be diagonal matrix written as follows
119876119894119894=
1maximum acceptable value of 1199092
119894
119877119895119895
=1
maximum acceptable value of 1199062119895
(32)
where
119894 = 120576 1 2 ℓ
119895 = 120576 1 2 119895 (33)
In essence Brysonrsquos rule scales the variables that appear in 119869 sothat the maximum acceptable value for each term is oneThisis especially important when the units used for the differentcomponents of 119906 and 119909 make the values for these variablesnumerically very different from each other
Assume that the steering ratio is 15 1 and the gains ofthe LQR controller are [01 098 0008] Figure 11 shows aresponse of the front axle system when the input steeringwheel angle is a step function as shown in Figure 11(a)
The results show that the angle of the front axle motorhas a fast response compared to the EPS steering system
International Journal of Vehicular Technology 9
0 2 4 6 8 100
2
4
6
8
10
12
14
16
18
20
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle
0 2 4 6 8 100
2
4
6
8
10
12
14
16
18
Time (s)
Mot
orc
olum
n an
gle (
deg)
EPS steeringmdashcolumn shaft angle
SBW systemmdashmotor angle
(b) Front axle motor angle
0 2 4 6 8 10
Rack
disp
lace
men
t (m
)
SBW system
EPS steering
Time (s)
18
16
14
12
10
8
6
4
2
0
times10minus3
(c) Rack displacement
0 2 4 6 8 100
02
04
06
08
1
12
14Fr
ont t
ire an
gle (
deg)
SBW system
EPS steering
Time (s)
(d) Front tire angle
Figure 11 Front axle system response for wheel synchronization
Moreover the LQR controller is able to reduce the delaycaused by the self aligning torque created from the contactbetween the tire and road surfaces This not only improvedthe vehicle stability but also ensured that the driver has amore confident level during manoeuvre Compared to theEPS steering system the delay caused by mechanical columnshaft connected to the steering wheel system Furthermore itcan be seen that the front tire angle is at almost fairly close to1∘ as shown in Figure 11(d) However it has small chatteringdue to the self aligning torque However with wide rangeadjustment gains of the LQR controller the steering responsecould be improved
32 Gain Scheduling (GS) with LQR Controller In certainsituations the dynamic behaviour of the process changed inaccordance with the conditions of the process This situation
could affect the stability of the system if not to control prop-erly Thus it is possible to change the gains of the controllerby monitoring the operating conditions of the process tohave a stable system The gain scheduling (GS) techniqueis used to change the gain of the LQR controller in order tocontrol the feedback torque and the steering feel gain (119896feel)is used to increase and decrease the torque It is important toincrease the feedback torque at high speed for steering stabil-ity and to decrease it in low speed especially during parkingto improve vehicle manoeuvre [9] This is based on the steer-ingwheel angle and vehicle speed function in order to achievethe feedback torque for the real driver steering feel
The relationship between a gain scheduling with LQRcontroller is shown in Table 4 The vehicle speed is dividedinto 3 categories at low speed (0ndash30 kmh) medium speed(31ndash100 kmh) and high speed (101ndash120 kmh) The maxi-mum vehicle speed is limited to 120 kmhThe steering wheel
10 International Journal of Vehicular Technology
0 2 4 6 8 10minus50
minus40
minus30
minus20
minus10
0
10
20
30
40
50
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Input steering wheel angle
0 1 2 3 4 5 6 7 8 9 10minus15
minus10
minus5
0
5
10
Time (s)
Stee
ring
torq
ue (N
m) LQR
LQR + GS
Reference
(b) 120583 = 1
0 1 2 3 4 5 6 7 8 9 10
minus15
minus10
minus5
0
5
10
15
Stee
ring
torq
ue (N
m) LQR
Reference
Time (s)
LQR + GS
(c) 120583 = 14
0 1 2 3 4 5 6 7 8 9 10minus20
minus15
minus10
minus5
0
5
10
15
20
Stee
ring
torq
ue (N
m) LQR LQR + GS
Reference
Time (s)
(d) 120583 = 17
Figure 12 Comparison of torque response of LQR and LQR + GS controller at different 120583
Table 4 The gains of LQR controller and steering feel
119881 (kmh) 120575119904119908
(degree) 1198961 1198962 1198963 119896feel
(0ndash30) (plusmn180∘ lt 120575119904119908
lt plusmn405∘) 003 065 0001 014(31ndash100) (plusmn46∘ lt 120575
119904119908lt plusmn179∘) 003 084 0001 060
(101ndash120) 120575119904119908
le +45∘ or 120575119904119908
le minus45∘ 005 095 0001 095
angle is also divided into 3 categories according to vehiclespeed condition The notation ldquo+rdquo means that the driverturned the steering wheel clockwise and notation ldquominusrdquo is usedfor counter clockwise in which the maximum steering wheelturn is plusmn405∘
The gains of the LQR controller are defined using Brysonrsquosrule method and the steering feel gain is defined usingintuitive and has a range from 01 to 1 This range has beendefined for various speeds 10 kmh 80 kmh and 120 kmhafter comparing with EPS steering system [15] For exampleif the driver turns a steering wheel between (plusmn180∘ lt 120575
119904119908lt
plusmn405∘) and vehicle speed is 10 kmh the total feedback torque
is multiplied by steering feel gain 119896feel = 014 Thus thefeedback torque is decreased which results in making thedriver feel it easier turning the steering wheel and improvesvehicle manoeuvre The feedback torque is regulated by LQRcontroller when the gains are 1198961 = 003 1198962 = 065 and1198963 = 0001 The gains are the optimum that has been assuredby providing a better torque control based on vehicle speedand steering wheel angle function
Comparing torque response between LQR and LQR +GScontroller for different coefficient (120583) is shown in Figure 12when the input steering wheel angle is a double lane changemanoeuvre as shown in Figure 12(a) The vehicle speed isset to be 80 kmh Based on results obtained the LQR + GStechnique offers an improvement 90 and the curve trend isalmost similar to reference torque compared to the LQR con-troller This is because the gains of LQR controller changedbased on the steering wheel angle and vehicle speed param-eter which is able to reject the uncertainty torque of a roadcondition Meanwhile with the fixed gains of the LQR con-troller it has an excessive torque that can cause the system tobe unstable The RMS torque values are illustrated in Table 5
International Journal of Vehicular Technology 11
Table 5 Comparison of RMS torque values of LQR and LQR + GScontroller
Reference LQR LQR + GS120583 = 1 56361 50795 55450120583 = 14 61218 68339 57243120583 = 17 66184 72888 62080
minus50 0 50minus15
minus10
minus5
0
5
10
15
Steering angle (deg)
Stee
ring
torq
ue (N
m) Feel gain = 5
Feel gain = 01Feel gain = 05
Feel gain = 1
(a) Different steering feel gain at 80 kmh
minus50 minus40 minus30 minus20 minus10 0 10 20 30 40 50
minus15
minus10
minus5
0
5
10
15
Steering wheel angle (deg)
Stee
ring
torq
ue (N
m)
80kmh kfeel = 060
120 kmh kfeel = 095
10kmh kfeel = 014
(b) Different steering feel gain at various speed
Figure 13 Steering feel gains response at various vehicle speed
Figure 13(a) shows the steering torque with respect tosteering wheel angle when different steering feel gain is usedat various vehicle speeds It is observed that when the steeringfeel gain is increased at maximum range which is 1 thesteering torque will increase and driver will sense stiffer onthe steering wheel at high speed manoeuvre Meanwhilewhen decreasing the steering feel gain this will result inreducing the steering torqueThus the driver will sense softer
Target monitor
Host PC
Cross over cableConnector block
PCI 6221
Power supplyCurrent sensor
DC motor driver
Figure 14 Steering wheel system hardware overview
during turning the steering wheel at the low speed manoeu-vre However when the steering feel gain is 5 which exceedsits maximum limit the steering torque becomes unstableFigure 13(b) shows the steering feel gains at various speedsAt high speed of 120 kmh the steering feel gain is increasedand this will increase the steering torque for steering stabilityMeanwhile at low speed the steering feel gain is decreasedwhich will help improve vehicle manoeuvre especially duringparking Based on the proposed gain scheduling with theLQR controller not only does it provide better control on thesteering torque but also it is optional for a designer to have awide range adjustment to vary the steering torque for a betterdriver steering feel and to improve vehicle manoeuvre
4 Hardware in the Loop (HIL) and Result
For hardware in the loop (HIL) setup the front axle systemand linear vehicle model are simulated on the host computerThe steering wheel system is replaced by HIL mechanismThe position of the steering wheel motor is measured usinga rotary encoder with 500 pulses per revolution To measurethe steering wheel torque the current sensor is connectedin series with the DC motor driver The steering wheel DCmotor has a torque 129Nm with 100 rpm The single tracklinear vehicle model is used to generate the estimation selfaligning torque as element to the feedback torque and distur-bance to the front axle system The HIL system is executedunder 1ms sampling time and interfaced using Matlab-XPCtarget toolbox configuration as shown in Figures 14 and 15
The assumptions through the experiment are listed asfollows
(i) the vehicle cruises at constant speed(ii) the condition of dry asphalt road (120583 = 085)
The experiments are done under high medium and lowspeed manoeuvre [26] The same input steering wheel anglebetween simulation and experiment is given Therefore theresult obtained from real time control is compared with theelectronic power steering (EPS) system [15]
41 Motor Model Validation for Steering Wheel SystemAlthough the model of steering wheel system has been
12 International Journal of Vehicular Technology
Stee
ring
whe
el sy
stem
DC motor
Current
Encoder
16 bit card PCI 6221
Analogoutput
Analoginput
Counterchannel
TargetPC
Monitortarget
PC
Crossover cable
Host PC
Front axle systemand
linear vehicle model
Hardware part Simulation part
Figure 15 Block diagram of the steering wheel-XPC target interfaced
Steering wheel model
120575m1
120575m1
ia1
r(t)
Figure 16 Block diagrammotormodel validation for steeringwheelsystem
derived the model validation is necessary to ensure thesystem is fairly able to adequate for real time applicationThis section describes the motor model validation of steeringwheel system The block diagram is shown in Figure 16
The input signal is 119903(119905) and the output is rate change ofmotor ( 120575
1198981) motor position (1205751198981) and the motor current
( 1198941198981) The measured output response is then used and com-
pared with simulation to estimate the parameter of the steer-ing wheel model using the Matlab tool software The mea-sured output data are taken based on the open loop systemwhen the input to the system is sinusoidal signal withamplitude of plusmn45∘ for 10 sec The Gaussian method is usedto estimate the parameters of the steering wheel motorThe comparison between simulation and measured outputis shown in Figure 17 It is obtained that the output of thesimulation is almost identical to the measured output How-ever small difference errors occur due to the small parametricuncertainties that have not been modelled The parameterestimation value is illustrated in Table 6
42 Noise Filtering Both analog and digital devices have atrait of noise or unwanted features that have effect on thesystem performance such as electromagnetic interferenceThe noise could be either random or white noise Attenuationof this noise is often as a primary goal in control systemapplication and in particular when controlling the processsystem The average value of noise typically is zero whichwill give misreading information to the process and it wouldnot be possible to control the process In order to reduce thenoise inference filtering technique is used A filter is used toremove unwanted noise It removes some noise frequenciesto suppress interfacing signal and reduce background noise
Table 6 Parameter estimation for DCmotor steering wheel system
Parameter Value Estimation value Unit1198771 4405 564 ohm1198711 0010 0017 henry119870119904119898
1459 0024 Nm1198691198981 0002 00036 (kgm2)1198871198981 0004 00068 Nm(rads)1198961199041 165 0025 Nm
There are many types of filtering techniques such as low passfilter that has been used in this study The low pass filter isused to filter out the interference noise from the output of thecurrent sensor
Figure 18 shows the comparison of the output responsefor a current sensor with and without using filters Withouta filter the current sensor has high amplitude interferencenoise However with low pass filter under the frequency of5Hz the interference noise is reduced and the fluctuations areremoved Thus this provides a smoother output response
43 Effect on Compensation Torque A steering torque com-parison without and with compensation torque is shown inFigure 19The results for without compensation torque showsthat the steering torque response curve is far away fromEPS steering system with different amplitude as shown inFigure 19(a) This happens due to less of derivative reactiontorque from the steering wheel motor Figure 19(b) showssteering torque response with compensation torque
It is observed that by adding the additional compensationtorque the steering torque response curve trends fairly closeto EPS steering system Thus it helps to improve the driversteering feel during manoeuvre and stabilize the systemHowever with the wide range adjustment gains of the com-pensation torque it could improve the steering torqueresponse
44 Medium Speed Manoeuvre The experimental steeringtorque results at medium speed manoeuvre are shown inFigure 20(b) when the steering wheel angle is plusmn100∘ as shown
International Journal of Vehicular Technology 13
0 2 4 6 8 10minus50
0
50
Time (s)
Rate
chan
ge m
otor
pos
ition
(deg
s)
MeasurementSimulation
(a) Rate change of motor position
0 2 4 6 8 100
10
20
30
40
50
60
Time (s)
Mot
or p
ositi
on (d
eg)
Measurement
Simulation
(b) Position of the motor
0 2 4 6 8 10minus01
minus005
0
005
01
015
Mot
or cu
rren
t (A
)
MeasurementSimulation
Time (s)
(c) Current of the motor
Figure 17 Motor model validation comparison between simulation and measurement
With filter Without filter
Time (s)
Curr
ent (
A)
008
006
004
002
0
minus002
minus004
minus006
minus0081 15 2 25 3 35 4 45 5 55
Figure 18 Noise filtering using low pass filter
in Figure 20(a) From the results the experimental steeringtorque response is almost identical to the EPS steering systemHowever it has small different torque amplitudes These
occur due to the fact that mechanism or friction from themechanical gearbox motor is not modelled in proposedcontrol algorithm
14 International Journal of Vehicular Technology
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
minus152 3 4 5 6 7 8 9 10
(a) Without compensation torque
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
21 3 4 5 6 7 8 9 10
(b) With compensation torque
Figure 19 Steering torque influence on compensation torque
minus80
minus60
minus40
minus20
0
20
40
60
80
100
Stee
ring
whe
el an
gle (
deg)
Time (s)3 4 5 6 7 8 9 10
(a) Steering wheel angle at 80 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
15
20
5
0
minus5
minus10
minus15
minus203 4 5 6 7 8 9 10
(b) Steering torque estimation at 80 kmh
Figure 20 Steering torque estimation and steering wheel angle response at 80 kmh
45 High Speed Manoeuvre At high speed manoeuvre theinput steering wheel angle that is more than plusmn45∘ is very dan-gerous [12] Therefore in this experiment the input steeringwheel angle is less than plusmn45∘ as shown in Figure 21(a) Basedon the results shown in Figures 21(b) and 21(c) the experi-mental steering torque is obtained fairly similar to the EPSsteering system Although the steering wheel angle is at smallangles the influence of vehicle speed cause the steering torqueis increased Moreover at higher speed the vehicle has moreunder steer propensity [9] This increased the self aligningtorque in proportional high torque at front axle motorThusthe driver will sense stiffer on the steering wheel duringmanoeuvre
In general the centering of the steering wheel perfor-mance or steering wheel returnability is better when thesteering torque gradient is large and the hysteresis is small [9]Based on this fact the result is shown in Figure 21(c) whereby
the hysteresis of the steering wheel returnability based onexperimental result is smaller compared to EPS steeringsystem In addition the steering torque gradient is larger thanEPS steering torque Therefore it is considered that the pro-posed feedback torque algorithmnot only achieved a steeringfeel but also improved the steering wheel returnability
46 Lower Speed Manoeuvre The steering torque results atlow speed manoeuvre are shown in Figures 22(b) and 22(c)when the input steering wheel is as shown in Figure 22(a)The vehicle cruised at constant speed of 10 kmh
The experimental steering torque is almost comparableto EPS torque but the amplitude is decreases The torque isdecreased due to the fact that the vehicle speed is reducedwhich is effect on reducing the demand to the total feedbacktorque Moreover the large turning of the steering wheelincreases the backlash of the motor which then contributes
International Journal of Vehicular Technology 15
2 3 4 5 6 7 8 9 10minus30
minus20
minus10
0
10
20
30
40
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle at 120 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steeringSimulationExperiment
10
5
0
minus5
minus10
3 4 5 6 7 8 9
(b) Steering torque estimation at 120 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment10
15
5
0
minus5
minus10
minus15minus20 minus10 0 10 20 30 40
Steering wheel angle (deg)
(c) Steering torque estimation against steering wheel angle at 120 kmh
Figure 21 Steering torque estimation response at 120 kmh
to the decrease of the motor torque Thus driver will sensesofter during steering the wheel during maneuver Howeverit has small different torque against EPS steering Wide rangeadjustment of gains in proposed control algorithm may helpimprove the steering torque response
3 4 5 6 7 8 9 10minus150
minus100
minus50
0
50
100
150
200
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
10
15
5
0
minus5
minus10
minus15
Time (s)2 3 4 5 6 7 8 9 10
(b) Steering torque estimation at 10 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
Steering wheel angle (deg)
10
8
6
4
2
0
minus2
minus4
minus6
minus8
minus10
minus100 minus50 0 50 100 150
(c) Steering torque estimation against steering wheel angle at 10 kmh
Figure 22 Steering torque estimation response at 10 kmh
5 Conclusion and Future Work
The proposed force feedback torque estimation and controlalgorithm have been described The proposed control algo-rithm is considered as an effort to generate the feedback
16 International Journal of Vehicular Technology
torque for driver steering feel in the vehicle SBW systemThe mathematical model of SBW subsystem consists ofsteering wheel and front axle systems that are modelledThe effectiveness of the control algorithm is verified throughthe experimental hardware in the loop (HIL) interfacedwith Matlab XPC target toolbox software According to theexperimental results the proposed control algorithm is ableto generate the feedback torque and the driver is able to sensesteering feel similar to that obtained in EPS steering systemThe LQR controller with gain scheduling based on steeringwheel angle and vehicle speed function provides better torquecontrol that allows rejection of the uncertainty torque froma road condition Furthermore the steering feel gain is ableto increase and lower the steering torque with improvedsteering stability at high speed and it improved the vehiclemanoeuvre at low speed It is also found that compensationtorque is able to improve the feedback torque and to stabilizethe system However it has small different torque magnitudethat occurs due to the backlash of steering wheel motor anduncertainty disturbance of mechanical properties Howeverwith a wide range adjustment of the gains on the proposedcontrol algorithm it may improve and provide realisticsteering torque for better driver steering feel
In future work the proposed control algorithm will betested in real vehicle in order to have more realistic steeringtorque for further investigation and analysis
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
Acknowledgments
This work is fully supported by the Ministry of Higher Edu-cation Malaysia and University Technology Malaysia underFRGS (vote 4F370) and Research University Grant (vote00G64) This work is also supported by Proton HoldingBerhad
References
[1] Y Yao ldquoVehicle steer-by-wire system controlrdquo SAE TechnicalPaper 2006-01-1175 2006
[2] E A Kumar D Dinesh and N Kamble ldquoAn overview of activefront steering systemrdquo International Journal of Scientific ampEngineering Research vol 3 no 6 2012
[3] S M H Fahami H Zamzuri S A Mazlan and N B Zulkar-nain ldquoThe design of vehicle active front steering based on steerby wire systemrdquo Advanced Science Letters vol 19 no 1 pp 61ndash65 2013
[4] S M H Fahami H Zamzuri S A Mazlan and S A SaruchildquoThe variable steering ratio for vehicle steer by wire systemusing hyperbolic tangent methodrdquo Applied Mechanics andMaterials vol 575 pp 781ndash784 2014
[5] T Kaufmann S Milsap B Murray and J Petrowski ldquoDevelop-ment experience with steer by wirerdquo in Proceedings of the SAEInternational Congress and Exhibition SAE 2001-01-2479 2001
[6] S Amberkare F Bolourchi J Demerly and S Millsap ldquoAcontrol system methodology for steer by wire systemsrdquo SAETechnical Paper 2004-01-1106 2004
[7] A Baviskar J R Wagner D M Dawson D Braganza andP Setlur ldquoAn adjustable steer-by-wire haptic-interface trackingcontroller for ground vehiclesrdquo IEEE Transactions on VehicularTechnology vol 58 no 2 pp 546ndash554 2009
[8] M Segawa S Kimura T Kada and S Nakona ldquoA study of reac-tive torque control for steer bywire systemrdquo inProceedings of theInternational Symposium on Advanced Vehicle Control (AVECrsquo02) pp 653ndash658 Hiroshima Japan September 2002
[9] S W Oh H C Chae S C Yun and C S Han ldquoThe design ofa controller for the steer-by-wire systemrdquo JSME InternationalJournal Series C Mechanical Systems Machine Elements andManufacturing vol 47 no 3 pp 896ndash907 2004
[10] C J Kim J H Jang S K Oh J Y Lee and J K HedrickldquoDevelopment of a control algorithm for a rack-actuating steer-by-wire system using road information feedbackrdquo Proceedingsof the Institution of Mechanical Engineers Part D Journal ofAutomobile Engineering vol 222 no 9 pp 1559ndash1571 2008
[11] S Asai H Kuroyanagi and S Takeuchi ldquoDevelopment of asteer-by-wire system with force feedback using a disturbanceobserverrdquo inProceedings of the SAEWorldCongressampExhibitionon Steering amp Suspension Technology Detroit Mich USA 2004
[12] T J Park C S Han and S H Lee ldquoDevelopment of the elec-tronic control unit for the rack-actuating steer-by-wire usingthe hardware-in-the-loop simulation systemrdquo Mechatronicsvol 15 no 8 pp 899ndash918 2005
[13] D Odenhtal T Bunte H D Heitzer and C Eicker ldquoHow tomake steer-by-wire feel like power steeringrdquo in Proceedings ofthe 15th IFACWorld Congress Barcelona Spain July 2000
[14] H Na C Zong and D Hu ldquoInvestigations on corneringcontrol algorithm design and road feeling optimization for asteer-by-wire vehiclerdquo in Proceedings of the IEEE InternationalConference on Mechatronics and Automation (ICMA rsquo09) pp3246ndash3251 Changchun China August 2009
[15] S Ancha A Baviskar J RWagner andDMDawson ldquoGroundvehicle steering systems modelling control and analysis ofhydraulic electric and steer-by-wire configurationsrdquo Interna-tional Journal of Vehicle Design vol 44 no 1-2 pp 188ndash2082007
[16] P Yih and J C Gerdes ldquoModification of vehicle handlingcharacteristics via steer-by-wirerdquo IEEE Transactions on ControlSystems Technology vol 13 no 6 pp 965ndash976 2005
[17] S M H Fahami H Zamzuri S A Mazlan and M A ZakarialdquoModeling and simulation of vehicle steer by wire systemrdquoin Proceedings of the IEEE Symposium on Humanities Scienceand Engineering Research (SHUSER rsquo12) pp 765ndash770 KualaLumpur Malaysia June 2012
[18] S Anwar and L Chen ldquoAn analytical redundancy-based faultdetection and isolation algorithm for a road-wheel controlsubsystem in a steer-by-wire systemrdquo IEEE Transactions onVehicular Technology vol 56 no 5 pp 2859ndash2869 2007
[19] M T Do Z Man C Zhang H Wang and F S Tay ldquoRobustsliding mode-based learning control for steer-by-wire systemsin modern vehiclesrdquo IEEE Transactions on Vehicular Technol-ogy vol 63 no 2 pp 580ndash590 2014
[20] J Duan R Wang and Y Yu ldquoResearch on control strategies ofsteer-by-wire systemrdquo in Proceedings of the International Con-ference on Intelligent Computation Technology and Automation(ICICTA rsquo10) vol 2 pp 1122ndash1125 May 2010
International Journal of Vehicular Technology 17
[21] B Zheng and S Anwar ldquoYaw stability control of a steer-by-wireequipped vehicle via active front wheel steeringrdquoMechatronicsvol 19 no 6 pp 799ndash804 2009
[22] S Sahoo S C Subramanian and S Srivastava ldquoDesign andimplementation of a controller for navigating an autonomousground vehiclerdquo in Proceedings of the 2nd International Confer-ence on Power Control and Embedded Systems (ICPCES rsquo12) pp1ndash6 IEEE Allahabad India December 2012
[23] T D Gillespie Fundamentals of Vehicle Dynamics Society ofAutomotive Engineers 1992
[24] M H Lee S K Ha J Y Choi and K S Yoon ldquoImprovement ofthe steering feel of an electric power steering system by torquemap modificationrdquo Journal of Mechanical Science and Technol-ogy vol 19 no 3 pp 792ndash801 2005
[25] R Pastorino M A Naya J A Perez and J Cuadrado ldquoGearedPM coreless motor modelling for driverrsquos force feedback insteer-by-wire systemsrdquo Mechatronics vol 21 no 6 pp 1043ndash1054 2011
[26] P Koehn andM Eckrich ldquoActive steeringmdashthe BMWapproachtowards modern steering technologyrdquo SAE Technical PaperSeries 2004-01-1105 SAE International 2004
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
6 International Journal of Vehicular Technology
+
+
+
minus
minus
Wheel synchronization
120575sw
120575sw
120591a
Frontaxle
system
LQRcontroller
LQRcontroller
120575m2
120575m2
120575f
ia2
120573
r
Linearvehiclemodel
Force feedback torque
120591map
120591fm
120591sm
120591damp
120591inertia
VGain
scheduling
kfm
kfeelsum
120591total feedback Steeringwheelsystem
Steering wheel system
Steeringwheel
120575m1
120575m1
ia1
ksm
Estimation selfaligning torque
Torque(RMS)
Figure 5 Overview of force feedback torque control algorithm for SBW system
Table 3 Linear vehicle model [16]
Parameter Description Value Unit119862119891 Front cornering stiffness 40000 Nrad
119862119903 Rear cornering stiffness 35000 Nrad
119897119891
Length from center frontwheel to (COG) vehicle 14 m
119897119903
Length from center rearwheel to COG vehicle 1 m
119898 Mass of the vehicle 1535 kg119868119904 Inertia of vehicle 2149 kgm2
119881 Speed of vehicle mdash kmh119865119910119891 Front tire lateral force mdash Nm
119865119910119903 Rear tire lateral force mdash Nm
120572119891 Front tire slip angle mdash Nm
120572119903 Rear tire slip angle mdash Nm
torque map is based on the function of the steering wheelangle and vehicle speed is added to the total feedback torqueMoreover a steering feel gain (119896feel) is adapted based on thefunction of steering wheel angle and vehicle speed to varythe feedback torque In order to control the feedback torquethe gain scheduling with LQR controller is used To showthe influence of the tire on the road surface the self aligningtorque (120591
119886) acts as input disturbance to the front axle system
as follows
120591119886= minus119862
119891(119905119901+ 119905119898) [120573+
119897119891119903
119881minus 120575119891]120583 (23)
where 120583 is coefficient for road dry condition pneumatic trail(119905119901) and mechanical trail (119905
119898)
On the other hand any change of the road surface couldchange the feedback torque magnitude of the front axlemotor Furthermore the LQR controller of the front axle sys-tem is used to synchronize the front tire angle (120575
119891) with driver
input steering wheel angle (120575119904119908) by controlling the angle of
the front axle motor (1205751198982) in accordance with the steering
ratioThus based on the proposed control algorithm the total
force feedback torque to the steering wheel system is writtenas follows
120591total feedback = ([(radicmean (120591119904119898
+ 120591119891119898
+ 120591map))]
+ 120591inertia + 120591damp) 119896feel
(24)
Author [9 10] has generated and discussed the force feed-back torque using the torque map conceptThe torque map isbased on the function of the steering wheel angle and vehiclespeed parameter The feedback torque is determined by twocontrol parameters (119896
120573) for the vehicle speed and (119896
120572) for the
steering wheel angle functionThe torque based on the function of vehicle speed is given
as follows [9]
1205911 = 1198961205731199092(13119909minus
12119881max)+119879in (25)
International Journal of Vehicular Technology 7
0 2 4 6 8 10minus40
minus30
minus20
minus10
0
10
20
30
40
Time (s)
Stee
ring
torq
ue (N
m)
70kmh75kmh
50kmh40kmh30kmh20kmh10kmh
80kmh
60kmh
Figure 6 Steering torque based on torque map at different speed
where (119909) is initial velocity and (119881max) is maximum velocitywhile the torque based on a function of steering wheel angleis written as follows [9]
1205912 = 119896120572radic1003816100381610038161003816120575119904119908
1003816100381610038161003816 sgn 120575119904119908 (26)
Based on the torque map concept the results of steeringtorque at different vehicle speed are shown in Figure 6 whenthe input steering wheel angle is a double lane changemanoeuvreThe torque increased at high speed anddecreasedat low speed [9 24] By varying the two control parametersthe steering reactive torque changes at different torque mag-nitude However the feedback torque based on the torquemap concept did not create the artificial steering feel dueto the fact that no element of road contact such as vehicledynamic responses are feedback to the torque map modelThus driver not sense realistic driver steering feel
Thus for this reason the front axle motor torque andsteering wheel motor torque are added to the torque mapThese ensure that the driver senses the feedback torque fromroad contact The front axle motor torque is chosen due toits proportion to the road surface The reaction torque of theroad surface gives the effect to the rack and pinion systemand it connects to the front axle DC motor Thus the frontaxle motor torque would always be proportional to the roadsurface Any changes in road surface will influence the frontaxle motor torque In order to measure the torque at the frontaxle motor the direct currentmeasurement technique is usedto estimate the torque
The direct relationship between torque and current of amotor can be defined [25]Thus then front axlemotor torque(120591119891119898
) that consists of the front axle motor constant (119896119891119898
) andthe motor current (119894
1198862) is written as follows
120591119891119898
= 1198941198862119896119891119898 (27)
and the corresponding steeringwheelmotor torque (120591119904119898) that
consists of the steering wheel motor constant (119896119904119898) and the
motor current (1198941198861) is written as follows
120591119904119898
= 1198941198861119896119904119898 (28)
0 2 4 6 8 10Time (s)
Mot
or to
rque
(Nm
)
Front axle motor torque
Steering wheel motor torque6
4
2
0
minus2
minus4
minus6
Figure 7 Steering wheel and front axle motor torque at 80 kmh
Figure 7 shows the front axle motor and steering wheelmotor torque response when the input is a double lanechangemanoeuvreThe results show that the front axlemotorgives a high torque This is due to the fact that the inputdisturbance torque between the tire and road surface which isthe self aligning torque has an effect on the front axle motorIncreasing the vehicle speedwill increase the front axlemotortorque On the other hand the magnitude of steering wheelmotor torque is decreased due to the elimination of mechan-ical column shaft and is not affected by the self aligningtorque
The phase compensation torque consists of the inertiatorque and damping torque potential to recreate the artificialsteering feel of the conventional steering system Moreoverit is intended to adjust the steering feel reduce the vibrationand stabilize the system Due to this fact the element of com-pensation torque is taken in order to improve the feedbacktorque for a driver steering feel and to stabilize the system
The damping torque (120591damp) which is the rate change ofthe steering wheel motor angle ( 120575
1198981) and the damping gain(119896damp) is written as follows
120591damp = 1205751198981119896damp (29)
The inertia torque (120591inertia) that is the acceleration of thesteering wheel motor angle ( 120575
1198981) and the inertia gain (119896inertia)is written as follows
120591inertia = 1205751198981119896inertia (30)
The inertia and damping torque responses are shown inFigure 8 when the input steering wheel angle is a double lanechange manoeuvre It is obtained that the damping torquegives a smooth signal and reduces the vibration Meanwhilethe inertia torque gives high magnitude Thus the combina-tion between the inertia torque and damping torque not onlyprovides an average of steering feel but is also able to stabilizethe system at the same time Figure 9 shows the feedback
8 International Journal of Vehicular Technology
0 2 4 6 8 10minus6
minus4
minus2
0
2
4
6
Time (s)
Torq
ue (N
m)
Inertia torque Damping torque
Figure 8 Inertia and damping torque
0 2 4 6 8 10Time (s)
Forc
e fee
dbac
k to
rque
(Nm
)
15
10
5
0
minus5
minus10
minus15
minus20
minus25
10kmh
50kmh100 kmh120 kmh
135kmh
Figure 9 Proposed force feedback torque response at differentvehicle speed
torque response based on (24) The torque is increasedaccording to the vehicle speed The higher vehicle speed willincrease the feedback torque and vice versa due to the samefact that is proposed by torque map
31 Linear Quadratic Regulator (LQR) Controller The LQRcontroller provides the best possible performance to controland stabilize the system by changing the location of the polesof the system to the optimal location for a time responseovershoot and steady stateThis is done by designing the statefeedback control119870 thatminimized cost function 119869 to stabilizethe system In this paper the LQR controller is used firstlyto control the front axle motor angle for the wheel synchro-nization and secondly to control the steering wheel motorfor a feedback torque
The wheel synchronization or directional control is abasic characteristic for steering function where a front tire
+
minus
sum
k1
k2
k3
x1
x2
x3
Plantr(t)
LQR controller
Figure 10 Basic control block diagram of LQR controller
angle should follow driver input steering wheel angle accord-ing to the steering ratio This is done by controlling theangle of the front axle motor In the LQR control schemesa feedback gain matrix is designed to achieve some compro-mise between the use of control effort the magnitude andthe speed of response that will guarantee a stable systemFigure 10 shows the control block diagram of the LQRcontroller
The optimal LQRproblem is often definedmore generallyand consists of finding controller transfer matrix that mini-mizes the cost function 119869 that is written as follows
119869 = int
prop
0(119909119879
119876119909+119906119879
119877119906) 119889119905 (31)
where 119909119879 and119880119879 respectively represent the transpose of the
state and input vector The 119876 and 119877 are positive definite (orpositive-semi definite) real symmetric matrix The 119876 and 119877
are usually chosen to be diagonal matrix written as follows
119876119894119894=
1maximum acceptable value of 1199092
119894
119877119895119895
=1
maximum acceptable value of 1199062119895
(32)
where
119894 = 120576 1 2 ℓ
119895 = 120576 1 2 119895 (33)
In essence Brysonrsquos rule scales the variables that appear in 119869 sothat the maximum acceptable value for each term is oneThisis especially important when the units used for the differentcomponents of 119906 and 119909 make the values for these variablesnumerically very different from each other
Assume that the steering ratio is 15 1 and the gains ofthe LQR controller are [01 098 0008] Figure 11 shows aresponse of the front axle system when the input steeringwheel angle is a step function as shown in Figure 11(a)
The results show that the angle of the front axle motorhas a fast response compared to the EPS steering system
International Journal of Vehicular Technology 9
0 2 4 6 8 100
2
4
6
8
10
12
14
16
18
20
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle
0 2 4 6 8 100
2
4
6
8
10
12
14
16
18
Time (s)
Mot
orc
olum
n an
gle (
deg)
EPS steeringmdashcolumn shaft angle
SBW systemmdashmotor angle
(b) Front axle motor angle
0 2 4 6 8 10
Rack
disp
lace
men
t (m
)
SBW system
EPS steering
Time (s)
18
16
14
12
10
8
6
4
2
0
times10minus3
(c) Rack displacement
0 2 4 6 8 100
02
04
06
08
1
12
14Fr
ont t
ire an
gle (
deg)
SBW system
EPS steering
Time (s)
(d) Front tire angle
Figure 11 Front axle system response for wheel synchronization
Moreover the LQR controller is able to reduce the delaycaused by the self aligning torque created from the contactbetween the tire and road surfaces This not only improvedthe vehicle stability but also ensured that the driver has amore confident level during manoeuvre Compared to theEPS steering system the delay caused by mechanical columnshaft connected to the steering wheel system Furthermore itcan be seen that the front tire angle is at almost fairly close to1∘ as shown in Figure 11(d) However it has small chatteringdue to the self aligning torque However with wide rangeadjustment gains of the LQR controller the steering responsecould be improved
32 Gain Scheduling (GS) with LQR Controller In certainsituations the dynamic behaviour of the process changed inaccordance with the conditions of the process This situation
could affect the stability of the system if not to control prop-erly Thus it is possible to change the gains of the controllerby monitoring the operating conditions of the process tohave a stable system The gain scheduling (GS) techniqueis used to change the gain of the LQR controller in order tocontrol the feedback torque and the steering feel gain (119896feel)is used to increase and decrease the torque It is important toincrease the feedback torque at high speed for steering stabil-ity and to decrease it in low speed especially during parkingto improve vehicle manoeuvre [9] This is based on the steer-ingwheel angle and vehicle speed function in order to achievethe feedback torque for the real driver steering feel
The relationship between a gain scheduling with LQRcontroller is shown in Table 4 The vehicle speed is dividedinto 3 categories at low speed (0ndash30 kmh) medium speed(31ndash100 kmh) and high speed (101ndash120 kmh) The maxi-mum vehicle speed is limited to 120 kmhThe steering wheel
10 International Journal of Vehicular Technology
0 2 4 6 8 10minus50
minus40
minus30
minus20
minus10
0
10
20
30
40
50
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Input steering wheel angle
0 1 2 3 4 5 6 7 8 9 10minus15
minus10
minus5
0
5
10
Time (s)
Stee
ring
torq
ue (N
m) LQR
LQR + GS
Reference
(b) 120583 = 1
0 1 2 3 4 5 6 7 8 9 10
minus15
minus10
minus5
0
5
10
15
Stee
ring
torq
ue (N
m) LQR
Reference
Time (s)
LQR + GS
(c) 120583 = 14
0 1 2 3 4 5 6 7 8 9 10minus20
minus15
minus10
minus5
0
5
10
15
20
Stee
ring
torq
ue (N
m) LQR LQR + GS
Reference
Time (s)
(d) 120583 = 17
Figure 12 Comparison of torque response of LQR and LQR + GS controller at different 120583
Table 4 The gains of LQR controller and steering feel
119881 (kmh) 120575119904119908
(degree) 1198961 1198962 1198963 119896feel
(0ndash30) (plusmn180∘ lt 120575119904119908
lt plusmn405∘) 003 065 0001 014(31ndash100) (plusmn46∘ lt 120575
119904119908lt plusmn179∘) 003 084 0001 060
(101ndash120) 120575119904119908
le +45∘ or 120575119904119908
le minus45∘ 005 095 0001 095
angle is also divided into 3 categories according to vehiclespeed condition The notation ldquo+rdquo means that the driverturned the steering wheel clockwise and notation ldquominusrdquo is usedfor counter clockwise in which the maximum steering wheelturn is plusmn405∘
The gains of the LQR controller are defined using Brysonrsquosrule method and the steering feel gain is defined usingintuitive and has a range from 01 to 1 This range has beendefined for various speeds 10 kmh 80 kmh and 120 kmhafter comparing with EPS steering system [15] For exampleif the driver turns a steering wheel between (plusmn180∘ lt 120575
119904119908lt
plusmn405∘) and vehicle speed is 10 kmh the total feedback torque
is multiplied by steering feel gain 119896feel = 014 Thus thefeedback torque is decreased which results in making thedriver feel it easier turning the steering wheel and improvesvehicle manoeuvre The feedback torque is regulated by LQRcontroller when the gains are 1198961 = 003 1198962 = 065 and1198963 = 0001 The gains are the optimum that has been assuredby providing a better torque control based on vehicle speedand steering wheel angle function
Comparing torque response between LQR and LQR +GScontroller for different coefficient (120583) is shown in Figure 12when the input steering wheel angle is a double lane changemanoeuvre as shown in Figure 12(a) The vehicle speed isset to be 80 kmh Based on results obtained the LQR + GStechnique offers an improvement 90 and the curve trend isalmost similar to reference torque compared to the LQR con-troller This is because the gains of LQR controller changedbased on the steering wheel angle and vehicle speed param-eter which is able to reject the uncertainty torque of a roadcondition Meanwhile with the fixed gains of the LQR con-troller it has an excessive torque that can cause the system tobe unstable The RMS torque values are illustrated in Table 5
International Journal of Vehicular Technology 11
Table 5 Comparison of RMS torque values of LQR and LQR + GScontroller
Reference LQR LQR + GS120583 = 1 56361 50795 55450120583 = 14 61218 68339 57243120583 = 17 66184 72888 62080
minus50 0 50minus15
minus10
minus5
0
5
10
15
Steering angle (deg)
Stee
ring
torq
ue (N
m) Feel gain = 5
Feel gain = 01Feel gain = 05
Feel gain = 1
(a) Different steering feel gain at 80 kmh
minus50 minus40 minus30 minus20 minus10 0 10 20 30 40 50
minus15
minus10
minus5
0
5
10
15
Steering wheel angle (deg)
Stee
ring
torq
ue (N
m)
80kmh kfeel = 060
120 kmh kfeel = 095
10kmh kfeel = 014
(b) Different steering feel gain at various speed
Figure 13 Steering feel gains response at various vehicle speed
Figure 13(a) shows the steering torque with respect tosteering wheel angle when different steering feel gain is usedat various vehicle speeds It is observed that when the steeringfeel gain is increased at maximum range which is 1 thesteering torque will increase and driver will sense stiffer onthe steering wheel at high speed manoeuvre Meanwhilewhen decreasing the steering feel gain this will result inreducing the steering torqueThus the driver will sense softer
Target monitor
Host PC
Cross over cableConnector block
PCI 6221
Power supplyCurrent sensor
DC motor driver
Figure 14 Steering wheel system hardware overview
during turning the steering wheel at the low speed manoeu-vre However when the steering feel gain is 5 which exceedsits maximum limit the steering torque becomes unstableFigure 13(b) shows the steering feel gains at various speedsAt high speed of 120 kmh the steering feel gain is increasedand this will increase the steering torque for steering stabilityMeanwhile at low speed the steering feel gain is decreasedwhich will help improve vehicle manoeuvre especially duringparking Based on the proposed gain scheduling with theLQR controller not only does it provide better control on thesteering torque but also it is optional for a designer to have awide range adjustment to vary the steering torque for a betterdriver steering feel and to improve vehicle manoeuvre
4 Hardware in the Loop (HIL) and Result
For hardware in the loop (HIL) setup the front axle systemand linear vehicle model are simulated on the host computerThe steering wheel system is replaced by HIL mechanismThe position of the steering wheel motor is measured usinga rotary encoder with 500 pulses per revolution To measurethe steering wheel torque the current sensor is connectedin series with the DC motor driver The steering wheel DCmotor has a torque 129Nm with 100 rpm The single tracklinear vehicle model is used to generate the estimation selfaligning torque as element to the feedback torque and distur-bance to the front axle system The HIL system is executedunder 1ms sampling time and interfaced using Matlab-XPCtarget toolbox configuration as shown in Figures 14 and 15
The assumptions through the experiment are listed asfollows
(i) the vehicle cruises at constant speed(ii) the condition of dry asphalt road (120583 = 085)
The experiments are done under high medium and lowspeed manoeuvre [26] The same input steering wheel anglebetween simulation and experiment is given Therefore theresult obtained from real time control is compared with theelectronic power steering (EPS) system [15]
41 Motor Model Validation for Steering Wheel SystemAlthough the model of steering wheel system has been
12 International Journal of Vehicular Technology
Stee
ring
whe
el sy
stem
DC motor
Current
Encoder
16 bit card PCI 6221
Analogoutput
Analoginput
Counterchannel
TargetPC
Monitortarget
PC
Crossover cable
Host PC
Front axle systemand
linear vehicle model
Hardware part Simulation part
Figure 15 Block diagram of the steering wheel-XPC target interfaced
Steering wheel model
120575m1
120575m1
ia1
r(t)
Figure 16 Block diagrammotormodel validation for steeringwheelsystem
derived the model validation is necessary to ensure thesystem is fairly able to adequate for real time applicationThis section describes the motor model validation of steeringwheel system The block diagram is shown in Figure 16
The input signal is 119903(119905) and the output is rate change ofmotor ( 120575
1198981) motor position (1205751198981) and the motor current
( 1198941198981) The measured output response is then used and com-
pared with simulation to estimate the parameter of the steer-ing wheel model using the Matlab tool software The mea-sured output data are taken based on the open loop systemwhen the input to the system is sinusoidal signal withamplitude of plusmn45∘ for 10 sec The Gaussian method is usedto estimate the parameters of the steering wheel motorThe comparison between simulation and measured outputis shown in Figure 17 It is obtained that the output of thesimulation is almost identical to the measured output How-ever small difference errors occur due to the small parametricuncertainties that have not been modelled The parameterestimation value is illustrated in Table 6
42 Noise Filtering Both analog and digital devices have atrait of noise or unwanted features that have effect on thesystem performance such as electromagnetic interferenceThe noise could be either random or white noise Attenuationof this noise is often as a primary goal in control systemapplication and in particular when controlling the processsystem The average value of noise typically is zero whichwill give misreading information to the process and it wouldnot be possible to control the process In order to reduce thenoise inference filtering technique is used A filter is used toremove unwanted noise It removes some noise frequenciesto suppress interfacing signal and reduce background noise
Table 6 Parameter estimation for DCmotor steering wheel system
Parameter Value Estimation value Unit1198771 4405 564 ohm1198711 0010 0017 henry119870119904119898
1459 0024 Nm1198691198981 0002 00036 (kgm2)1198871198981 0004 00068 Nm(rads)1198961199041 165 0025 Nm
There are many types of filtering techniques such as low passfilter that has been used in this study The low pass filter isused to filter out the interference noise from the output of thecurrent sensor
Figure 18 shows the comparison of the output responsefor a current sensor with and without using filters Withouta filter the current sensor has high amplitude interferencenoise However with low pass filter under the frequency of5Hz the interference noise is reduced and the fluctuations areremoved Thus this provides a smoother output response
43 Effect on Compensation Torque A steering torque com-parison without and with compensation torque is shown inFigure 19The results for without compensation torque showsthat the steering torque response curve is far away fromEPS steering system with different amplitude as shown inFigure 19(a) This happens due to less of derivative reactiontorque from the steering wheel motor Figure 19(b) showssteering torque response with compensation torque
It is observed that by adding the additional compensationtorque the steering torque response curve trends fairly closeto EPS steering system Thus it helps to improve the driversteering feel during manoeuvre and stabilize the systemHowever with the wide range adjustment gains of the com-pensation torque it could improve the steering torqueresponse
44 Medium Speed Manoeuvre The experimental steeringtorque results at medium speed manoeuvre are shown inFigure 20(b) when the steering wheel angle is plusmn100∘ as shown
International Journal of Vehicular Technology 13
0 2 4 6 8 10minus50
0
50
Time (s)
Rate
chan
ge m
otor
pos
ition
(deg
s)
MeasurementSimulation
(a) Rate change of motor position
0 2 4 6 8 100
10
20
30
40
50
60
Time (s)
Mot
or p
ositi
on (d
eg)
Measurement
Simulation
(b) Position of the motor
0 2 4 6 8 10minus01
minus005
0
005
01
015
Mot
or cu
rren
t (A
)
MeasurementSimulation
Time (s)
(c) Current of the motor
Figure 17 Motor model validation comparison between simulation and measurement
With filter Without filter
Time (s)
Curr
ent (
A)
008
006
004
002
0
minus002
minus004
minus006
minus0081 15 2 25 3 35 4 45 5 55
Figure 18 Noise filtering using low pass filter
in Figure 20(a) From the results the experimental steeringtorque response is almost identical to the EPS steering systemHowever it has small different torque amplitudes These
occur due to the fact that mechanism or friction from themechanical gearbox motor is not modelled in proposedcontrol algorithm
14 International Journal of Vehicular Technology
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
minus152 3 4 5 6 7 8 9 10
(a) Without compensation torque
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
21 3 4 5 6 7 8 9 10
(b) With compensation torque
Figure 19 Steering torque influence on compensation torque
minus80
minus60
minus40
minus20
0
20
40
60
80
100
Stee
ring
whe
el an
gle (
deg)
Time (s)3 4 5 6 7 8 9 10
(a) Steering wheel angle at 80 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
15
20
5
0
minus5
minus10
minus15
minus203 4 5 6 7 8 9 10
(b) Steering torque estimation at 80 kmh
Figure 20 Steering torque estimation and steering wheel angle response at 80 kmh
45 High Speed Manoeuvre At high speed manoeuvre theinput steering wheel angle that is more than plusmn45∘ is very dan-gerous [12] Therefore in this experiment the input steeringwheel angle is less than plusmn45∘ as shown in Figure 21(a) Basedon the results shown in Figures 21(b) and 21(c) the experi-mental steering torque is obtained fairly similar to the EPSsteering system Although the steering wheel angle is at smallangles the influence of vehicle speed cause the steering torqueis increased Moreover at higher speed the vehicle has moreunder steer propensity [9] This increased the self aligningtorque in proportional high torque at front axle motorThusthe driver will sense stiffer on the steering wheel duringmanoeuvre
In general the centering of the steering wheel perfor-mance or steering wheel returnability is better when thesteering torque gradient is large and the hysteresis is small [9]Based on this fact the result is shown in Figure 21(c) whereby
the hysteresis of the steering wheel returnability based onexperimental result is smaller compared to EPS steeringsystem In addition the steering torque gradient is larger thanEPS steering torque Therefore it is considered that the pro-posed feedback torque algorithmnot only achieved a steeringfeel but also improved the steering wheel returnability
46 Lower Speed Manoeuvre The steering torque results atlow speed manoeuvre are shown in Figures 22(b) and 22(c)when the input steering wheel is as shown in Figure 22(a)The vehicle cruised at constant speed of 10 kmh
The experimental steering torque is almost comparableto EPS torque but the amplitude is decreases The torque isdecreased due to the fact that the vehicle speed is reducedwhich is effect on reducing the demand to the total feedbacktorque Moreover the large turning of the steering wheelincreases the backlash of the motor which then contributes
International Journal of Vehicular Technology 15
2 3 4 5 6 7 8 9 10minus30
minus20
minus10
0
10
20
30
40
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle at 120 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steeringSimulationExperiment
10
5
0
minus5
minus10
3 4 5 6 7 8 9
(b) Steering torque estimation at 120 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment10
15
5
0
minus5
minus10
minus15minus20 minus10 0 10 20 30 40
Steering wheel angle (deg)
(c) Steering torque estimation against steering wheel angle at 120 kmh
Figure 21 Steering torque estimation response at 120 kmh
to the decrease of the motor torque Thus driver will sensesofter during steering the wheel during maneuver Howeverit has small different torque against EPS steering Wide rangeadjustment of gains in proposed control algorithm may helpimprove the steering torque response
3 4 5 6 7 8 9 10minus150
minus100
minus50
0
50
100
150
200
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
10
15
5
0
minus5
minus10
minus15
Time (s)2 3 4 5 6 7 8 9 10
(b) Steering torque estimation at 10 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
Steering wheel angle (deg)
10
8
6
4
2
0
minus2
minus4
minus6
minus8
minus10
minus100 minus50 0 50 100 150
(c) Steering torque estimation against steering wheel angle at 10 kmh
Figure 22 Steering torque estimation response at 10 kmh
5 Conclusion and Future Work
The proposed force feedback torque estimation and controlalgorithm have been described The proposed control algo-rithm is considered as an effort to generate the feedback
16 International Journal of Vehicular Technology
torque for driver steering feel in the vehicle SBW systemThe mathematical model of SBW subsystem consists ofsteering wheel and front axle systems that are modelledThe effectiveness of the control algorithm is verified throughthe experimental hardware in the loop (HIL) interfacedwith Matlab XPC target toolbox software According to theexperimental results the proposed control algorithm is ableto generate the feedback torque and the driver is able to sensesteering feel similar to that obtained in EPS steering systemThe LQR controller with gain scheduling based on steeringwheel angle and vehicle speed function provides better torquecontrol that allows rejection of the uncertainty torque froma road condition Furthermore the steering feel gain is ableto increase and lower the steering torque with improvedsteering stability at high speed and it improved the vehiclemanoeuvre at low speed It is also found that compensationtorque is able to improve the feedback torque and to stabilizethe system However it has small different torque magnitudethat occurs due to the backlash of steering wheel motor anduncertainty disturbance of mechanical properties Howeverwith a wide range adjustment of the gains on the proposedcontrol algorithm it may improve and provide realisticsteering torque for better driver steering feel
In future work the proposed control algorithm will betested in real vehicle in order to have more realistic steeringtorque for further investigation and analysis
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
Acknowledgments
This work is fully supported by the Ministry of Higher Edu-cation Malaysia and University Technology Malaysia underFRGS (vote 4F370) and Research University Grant (vote00G64) This work is also supported by Proton HoldingBerhad
References
[1] Y Yao ldquoVehicle steer-by-wire system controlrdquo SAE TechnicalPaper 2006-01-1175 2006
[2] E A Kumar D Dinesh and N Kamble ldquoAn overview of activefront steering systemrdquo International Journal of Scientific ampEngineering Research vol 3 no 6 2012
[3] S M H Fahami H Zamzuri S A Mazlan and N B Zulkar-nain ldquoThe design of vehicle active front steering based on steerby wire systemrdquo Advanced Science Letters vol 19 no 1 pp 61ndash65 2013
[4] S M H Fahami H Zamzuri S A Mazlan and S A SaruchildquoThe variable steering ratio for vehicle steer by wire systemusing hyperbolic tangent methodrdquo Applied Mechanics andMaterials vol 575 pp 781ndash784 2014
[5] T Kaufmann S Milsap B Murray and J Petrowski ldquoDevelop-ment experience with steer by wirerdquo in Proceedings of the SAEInternational Congress and Exhibition SAE 2001-01-2479 2001
[6] S Amberkare F Bolourchi J Demerly and S Millsap ldquoAcontrol system methodology for steer by wire systemsrdquo SAETechnical Paper 2004-01-1106 2004
[7] A Baviskar J R Wagner D M Dawson D Braganza andP Setlur ldquoAn adjustable steer-by-wire haptic-interface trackingcontroller for ground vehiclesrdquo IEEE Transactions on VehicularTechnology vol 58 no 2 pp 546ndash554 2009
[8] M Segawa S Kimura T Kada and S Nakona ldquoA study of reac-tive torque control for steer bywire systemrdquo inProceedings of theInternational Symposium on Advanced Vehicle Control (AVECrsquo02) pp 653ndash658 Hiroshima Japan September 2002
[9] S W Oh H C Chae S C Yun and C S Han ldquoThe design ofa controller for the steer-by-wire systemrdquo JSME InternationalJournal Series C Mechanical Systems Machine Elements andManufacturing vol 47 no 3 pp 896ndash907 2004
[10] C J Kim J H Jang S K Oh J Y Lee and J K HedrickldquoDevelopment of a control algorithm for a rack-actuating steer-by-wire system using road information feedbackrdquo Proceedingsof the Institution of Mechanical Engineers Part D Journal ofAutomobile Engineering vol 222 no 9 pp 1559ndash1571 2008
[11] S Asai H Kuroyanagi and S Takeuchi ldquoDevelopment of asteer-by-wire system with force feedback using a disturbanceobserverrdquo inProceedings of the SAEWorldCongressampExhibitionon Steering amp Suspension Technology Detroit Mich USA 2004
[12] T J Park C S Han and S H Lee ldquoDevelopment of the elec-tronic control unit for the rack-actuating steer-by-wire usingthe hardware-in-the-loop simulation systemrdquo Mechatronicsvol 15 no 8 pp 899ndash918 2005
[13] D Odenhtal T Bunte H D Heitzer and C Eicker ldquoHow tomake steer-by-wire feel like power steeringrdquo in Proceedings ofthe 15th IFACWorld Congress Barcelona Spain July 2000
[14] H Na C Zong and D Hu ldquoInvestigations on corneringcontrol algorithm design and road feeling optimization for asteer-by-wire vehiclerdquo in Proceedings of the IEEE InternationalConference on Mechatronics and Automation (ICMA rsquo09) pp3246ndash3251 Changchun China August 2009
[15] S Ancha A Baviskar J RWagner andDMDawson ldquoGroundvehicle steering systems modelling control and analysis ofhydraulic electric and steer-by-wire configurationsrdquo Interna-tional Journal of Vehicle Design vol 44 no 1-2 pp 188ndash2082007
[16] P Yih and J C Gerdes ldquoModification of vehicle handlingcharacteristics via steer-by-wirerdquo IEEE Transactions on ControlSystems Technology vol 13 no 6 pp 965ndash976 2005
[17] S M H Fahami H Zamzuri S A Mazlan and M A ZakarialdquoModeling and simulation of vehicle steer by wire systemrdquoin Proceedings of the IEEE Symposium on Humanities Scienceand Engineering Research (SHUSER rsquo12) pp 765ndash770 KualaLumpur Malaysia June 2012
[18] S Anwar and L Chen ldquoAn analytical redundancy-based faultdetection and isolation algorithm for a road-wheel controlsubsystem in a steer-by-wire systemrdquo IEEE Transactions onVehicular Technology vol 56 no 5 pp 2859ndash2869 2007
[19] M T Do Z Man C Zhang H Wang and F S Tay ldquoRobustsliding mode-based learning control for steer-by-wire systemsin modern vehiclesrdquo IEEE Transactions on Vehicular Technol-ogy vol 63 no 2 pp 580ndash590 2014
[20] J Duan R Wang and Y Yu ldquoResearch on control strategies ofsteer-by-wire systemrdquo in Proceedings of the International Con-ference on Intelligent Computation Technology and Automation(ICICTA rsquo10) vol 2 pp 1122ndash1125 May 2010
International Journal of Vehicular Technology 17
[21] B Zheng and S Anwar ldquoYaw stability control of a steer-by-wireequipped vehicle via active front wheel steeringrdquoMechatronicsvol 19 no 6 pp 799ndash804 2009
[22] S Sahoo S C Subramanian and S Srivastava ldquoDesign andimplementation of a controller for navigating an autonomousground vehiclerdquo in Proceedings of the 2nd International Confer-ence on Power Control and Embedded Systems (ICPCES rsquo12) pp1ndash6 IEEE Allahabad India December 2012
[23] T D Gillespie Fundamentals of Vehicle Dynamics Society ofAutomotive Engineers 1992
[24] M H Lee S K Ha J Y Choi and K S Yoon ldquoImprovement ofthe steering feel of an electric power steering system by torquemap modificationrdquo Journal of Mechanical Science and Technol-ogy vol 19 no 3 pp 792ndash801 2005
[25] R Pastorino M A Naya J A Perez and J Cuadrado ldquoGearedPM coreless motor modelling for driverrsquos force feedback insteer-by-wire systemsrdquo Mechatronics vol 21 no 6 pp 1043ndash1054 2011
[26] P Koehn andM Eckrich ldquoActive steeringmdashthe BMWapproachtowards modern steering technologyrdquo SAE Technical PaperSeries 2004-01-1105 SAE International 2004
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
International Journal of Vehicular Technology 7
0 2 4 6 8 10minus40
minus30
minus20
minus10
0
10
20
30
40
Time (s)
Stee
ring
torq
ue (N
m)
70kmh75kmh
50kmh40kmh30kmh20kmh10kmh
80kmh
60kmh
Figure 6 Steering torque based on torque map at different speed
where (119909) is initial velocity and (119881max) is maximum velocitywhile the torque based on a function of steering wheel angleis written as follows [9]
1205912 = 119896120572radic1003816100381610038161003816120575119904119908
1003816100381610038161003816 sgn 120575119904119908 (26)
Based on the torque map concept the results of steeringtorque at different vehicle speed are shown in Figure 6 whenthe input steering wheel angle is a double lane changemanoeuvreThe torque increased at high speed anddecreasedat low speed [9 24] By varying the two control parametersthe steering reactive torque changes at different torque mag-nitude However the feedback torque based on the torquemap concept did not create the artificial steering feel dueto the fact that no element of road contact such as vehicledynamic responses are feedback to the torque map modelThus driver not sense realistic driver steering feel
Thus for this reason the front axle motor torque andsteering wheel motor torque are added to the torque mapThese ensure that the driver senses the feedback torque fromroad contact The front axle motor torque is chosen due toits proportion to the road surface The reaction torque of theroad surface gives the effect to the rack and pinion systemand it connects to the front axle DC motor Thus the frontaxle motor torque would always be proportional to the roadsurface Any changes in road surface will influence the frontaxle motor torque In order to measure the torque at the frontaxle motor the direct currentmeasurement technique is usedto estimate the torque
The direct relationship between torque and current of amotor can be defined [25]Thus then front axlemotor torque(120591119891119898
) that consists of the front axle motor constant (119896119891119898
) andthe motor current (119894
1198862) is written as follows
120591119891119898
= 1198941198862119896119891119898 (27)
and the corresponding steeringwheelmotor torque (120591119904119898) that
consists of the steering wheel motor constant (119896119904119898) and the
motor current (1198941198861) is written as follows
120591119904119898
= 1198941198861119896119904119898 (28)
0 2 4 6 8 10Time (s)
Mot
or to
rque
(Nm
)
Front axle motor torque
Steering wheel motor torque6
4
2
0
minus2
minus4
minus6
Figure 7 Steering wheel and front axle motor torque at 80 kmh
Figure 7 shows the front axle motor and steering wheelmotor torque response when the input is a double lanechangemanoeuvreThe results show that the front axlemotorgives a high torque This is due to the fact that the inputdisturbance torque between the tire and road surface which isthe self aligning torque has an effect on the front axle motorIncreasing the vehicle speedwill increase the front axlemotortorque On the other hand the magnitude of steering wheelmotor torque is decreased due to the elimination of mechan-ical column shaft and is not affected by the self aligningtorque
The phase compensation torque consists of the inertiatorque and damping torque potential to recreate the artificialsteering feel of the conventional steering system Moreoverit is intended to adjust the steering feel reduce the vibrationand stabilize the system Due to this fact the element of com-pensation torque is taken in order to improve the feedbacktorque for a driver steering feel and to stabilize the system
The damping torque (120591damp) which is the rate change ofthe steering wheel motor angle ( 120575
1198981) and the damping gain(119896damp) is written as follows
120591damp = 1205751198981119896damp (29)
The inertia torque (120591inertia) that is the acceleration of thesteering wheel motor angle ( 120575
1198981) and the inertia gain (119896inertia)is written as follows
120591inertia = 1205751198981119896inertia (30)
The inertia and damping torque responses are shown inFigure 8 when the input steering wheel angle is a double lanechange manoeuvre It is obtained that the damping torquegives a smooth signal and reduces the vibration Meanwhilethe inertia torque gives high magnitude Thus the combina-tion between the inertia torque and damping torque not onlyprovides an average of steering feel but is also able to stabilizethe system at the same time Figure 9 shows the feedback
8 International Journal of Vehicular Technology
0 2 4 6 8 10minus6
minus4
minus2
0
2
4
6
Time (s)
Torq
ue (N
m)
Inertia torque Damping torque
Figure 8 Inertia and damping torque
0 2 4 6 8 10Time (s)
Forc
e fee
dbac
k to
rque
(Nm
)
15
10
5
0
minus5
minus10
minus15
minus20
minus25
10kmh
50kmh100 kmh120 kmh
135kmh
Figure 9 Proposed force feedback torque response at differentvehicle speed
torque response based on (24) The torque is increasedaccording to the vehicle speed The higher vehicle speed willincrease the feedback torque and vice versa due to the samefact that is proposed by torque map
31 Linear Quadratic Regulator (LQR) Controller The LQRcontroller provides the best possible performance to controland stabilize the system by changing the location of the polesof the system to the optimal location for a time responseovershoot and steady stateThis is done by designing the statefeedback control119870 thatminimized cost function 119869 to stabilizethe system In this paper the LQR controller is used firstlyto control the front axle motor angle for the wheel synchro-nization and secondly to control the steering wheel motorfor a feedback torque
The wheel synchronization or directional control is abasic characteristic for steering function where a front tire
+
minus
sum
k1
k2
k3
x1
x2
x3
Plantr(t)
LQR controller
Figure 10 Basic control block diagram of LQR controller
angle should follow driver input steering wheel angle accord-ing to the steering ratio This is done by controlling theangle of the front axle motor In the LQR control schemesa feedback gain matrix is designed to achieve some compro-mise between the use of control effort the magnitude andthe speed of response that will guarantee a stable systemFigure 10 shows the control block diagram of the LQRcontroller
The optimal LQRproblem is often definedmore generallyand consists of finding controller transfer matrix that mini-mizes the cost function 119869 that is written as follows
119869 = int
prop
0(119909119879
119876119909+119906119879
119877119906) 119889119905 (31)
where 119909119879 and119880119879 respectively represent the transpose of the
state and input vector The 119876 and 119877 are positive definite (orpositive-semi definite) real symmetric matrix The 119876 and 119877
are usually chosen to be diagonal matrix written as follows
119876119894119894=
1maximum acceptable value of 1199092
119894
119877119895119895
=1
maximum acceptable value of 1199062119895
(32)
where
119894 = 120576 1 2 ℓ
119895 = 120576 1 2 119895 (33)
In essence Brysonrsquos rule scales the variables that appear in 119869 sothat the maximum acceptable value for each term is oneThisis especially important when the units used for the differentcomponents of 119906 and 119909 make the values for these variablesnumerically very different from each other
Assume that the steering ratio is 15 1 and the gains ofthe LQR controller are [01 098 0008] Figure 11 shows aresponse of the front axle system when the input steeringwheel angle is a step function as shown in Figure 11(a)
The results show that the angle of the front axle motorhas a fast response compared to the EPS steering system
International Journal of Vehicular Technology 9
0 2 4 6 8 100
2
4
6
8
10
12
14
16
18
20
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle
0 2 4 6 8 100
2
4
6
8
10
12
14
16
18
Time (s)
Mot
orc
olum
n an
gle (
deg)
EPS steeringmdashcolumn shaft angle
SBW systemmdashmotor angle
(b) Front axle motor angle
0 2 4 6 8 10
Rack
disp
lace
men
t (m
)
SBW system
EPS steering
Time (s)
18
16
14
12
10
8
6
4
2
0
times10minus3
(c) Rack displacement
0 2 4 6 8 100
02
04
06
08
1
12
14Fr
ont t
ire an
gle (
deg)
SBW system
EPS steering
Time (s)
(d) Front tire angle
Figure 11 Front axle system response for wheel synchronization
Moreover the LQR controller is able to reduce the delaycaused by the self aligning torque created from the contactbetween the tire and road surfaces This not only improvedthe vehicle stability but also ensured that the driver has amore confident level during manoeuvre Compared to theEPS steering system the delay caused by mechanical columnshaft connected to the steering wheel system Furthermore itcan be seen that the front tire angle is at almost fairly close to1∘ as shown in Figure 11(d) However it has small chatteringdue to the self aligning torque However with wide rangeadjustment gains of the LQR controller the steering responsecould be improved
32 Gain Scheduling (GS) with LQR Controller In certainsituations the dynamic behaviour of the process changed inaccordance with the conditions of the process This situation
could affect the stability of the system if not to control prop-erly Thus it is possible to change the gains of the controllerby monitoring the operating conditions of the process tohave a stable system The gain scheduling (GS) techniqueis used to change the gain of the LQR controller in order tocontrol the feedback torque and the steering feel gain (119896feel)is used to increase and decrease the torque It is important toincrease the feedback torque at high speed for steering stabil-ity and to decrease it in low speed especially during parkingto improve vehicle manoeuvre [9] This is based on the steer-ingwheel angle and vehicle speed function in order to achievethe feedback torque for the real driver steering feel
The relationship between a gain scheduling with LQRcontroller is shown in Table 4 The vehicle speed is dividedinto 3 categories at low speed (0ndash30 kmh) medium speed(31ndash100 kmh) and high speed (101ndash120 kmh) The maxi-mum vehicle speed is limited to 120 kmhThe steering wheel
10 International Journal of Vehicular Technology
0 2 4 6 8 10minus50
minus40
minus30
minus20
minus10
0
10
20
30
40
50
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Input steering wheel angle
0 1 2 3 4 5 6 7 8 9 10minus15
minus10
minus5
0
5
10
Time (s)
Stee
ring
torq
ue (N
m) LQR
LQR + GS
Reference
(b) 120583 = 1
0 1 2 3 4 5 6 7 8 9 10
minus15
minus10
minus5
0
5
10
15
Stee
ring
torq
ue (N
m) LQR
Reference
Time (s)
LQR + GS
(c) 120583 = 14
0 1 2 3 4 5 6 7 8 9 10minus20
minus15
minus10
minus5
0
5
10
15
20
Stee
ring
torq
ue (N
m) LQR LQR + GS
Reference
Time (s)
(d) 120583 = 17
Figure 12 Comparison of torque response of LQR and LQR + GS controller at different 120583
Table 4 The gains of LQR controller and steering feel
119881 (kmh) 120575119904119908
(degree) 1198961 1198962 1198963 119896feel
(0ndash30) (plusmn180∘ lt 120575119904119908
lt plusmn405∘) 003 065 0001 014(31ndash100) (plusmn46∘ lt 120575
119904119908lt plusmn179∘) 003 084 0001 060
(101ndash120) 120575119904119908
le +45∘ or 120575119904119908
le minus45∘ 005 095 0001 095
angle is also divided into 3 categories according to vehiclespeed condition The notation ldquo+rdquo means that the driverturned the steering wheel clockwise and notation ldquominusrdquo is usedfor counter clockwise in which the maximum steering wheelturn is plusmn405∘
The gains of the LQR controller are defined using Brysonrsquosrule method and the steering feel gain is defined usingintuitive and has a range from 01 to 1 This range has beendefined for various speeds 10 kmh 80 kmh and 120 kmhafter comparing with EPS steering system [15] For exampleif the driver turns a steering wheel between (plusmn180∘ lt 120575
119904119908lt
plusmn405∘) and vehicle speed is 10 kmh the total feedback torque
is multiplied by steering feel gain 119896feel = 014 Thus thefeedback torque is decreased which results in making thedriver feel it easier turning the steering wheel and improvesvehicle manoeuvre The feedback torque is regulated by LQRcontroller when the gains are 1198961 = 003 1198962 = 065 and1198963 = 0001 The gains are the optimum that has been assuredby providing a better torque control based on vehicle speedand steering wheel angle function
Comparing torque response between LQR and LQR +GScontroller for different coefficient (120583) is shown in Figure 12when the input steering wheel angle is a double lane changemanoeuvre as shown in Figure 12(a) The vehicle speed isset to be 80 kmh Based on results obtained the LQR + GStechnique offers an improvement 90 and the curve trend isalmost similar to reference torque compared to the LQR con-troller This is because the gains of LQR controller changedbased on the steering wheel angle and vehicle speed param-eter which is able to reject the uncertainty torque of a roadcondition Meanwhile with the fixed gains of the LQR con-troller it has an excessive torque that can cause the system tobe unstable The RMS torque values are illustrated in Table 5
International Journal of Vehicular Technology 11
Table 5 Comparison of RMS torque values of LQR and LQR + GScontroller
Reference LQR LQR + GS120583 = 1 56361 50795 55450120583 = 14 61218 68339 57243120583 = 17 66184 72888 62080
minus50 0 50minus15
minus10
minus5
0
5
10
15
Steering angle (deg)
Stee
ring
torq
ue (N
m) Feel gain = 5
Feel gain = 01Feel gain = 05
Feel gain = 1
(a) Different steering feel gain at 80 kmh
minus50 minus40 minus30 minus20 minus10 0 10 20 30 40 50
minus15
minus10
minus5
0
5
10
15
Steering wheel angle (deg)
Stee
ring
torq
ue (N
m)
80kmh kfeel = 060
120 kmh kfeel = 095
10kmh kfeel = 014
(b) Different steering feel gain at various speed
Figure 13 Steering feel gains response at various vehicle speed
Figure 13(a) shows the steering torque with respect tosteering wheel angle when different steering feel gain is usedat various vehicle speeds It is observed that when the steeringfeel gain is increased at maximum range which is 1 thesteering torque will increase and driver will sense stiffer onthe steering wheel at high speed manoeuvre Meanwhilewhen decreasing the steering feel gain this will result inreducing the steering torqueThus the driver will sense softer
Target monitor
Host PC
Cross over cableConnector block
PCI 6221
Power supplyCurrent sensor
DC motor driver
Figure 14 Steering wheel system hardware overview
during turning the steering wheel at the low speed manoeu-vre However when the steering feel gain is 5 which exceedsits maximum limit the steering torque becomes unstableFigure 13(b) shows the steering feel gains at various speedsAt high speed of 120 kmh the steering feel gain is increasedand this will increase the steering torque for steering stabilityMeanwhile at low speed the steering feel gain is decreasedwhich will help improve vehicle manoeuvre especially duringparking Based on the proposed gain scheduling with theLQR controller not only does it provide better control on thesteering torque but also it is optional for a designer to have awide range adjustment to vary the steering torque for a betterdriver steering feel and to improve vehicle manoeuvre
4 Hardware in the Loop (HIL) and Result
For hardware in the loop (HIL) setup the front axle systemand linear vehicle model are simulated on the host computerThe steering wheel system is replaced by HIL mechanismThe position of the steering wheel motor is measured usinga rotary encoder with 500 pulses per revolution To measurethe steering wheel torque the current sensor is connectedin series with the DC motor driver The steering wheel DCmotor has a torque 129Nm with 100 rpm The single tracklinear vehicle model is used to generate the estimation selfaligning torque as element to the feedback torque and distur-bance to the front axle system The HIL system is executedunder 1ms sampling time and interfaced using Matlab-XPCtarget toolbox configuration as shown in Figures 14 and 15
The assumptions through the experiment are listed asfollows
(i) the vehicle cruises at constant speed(ii) the condition of dry asphalt road (120583 = 085)
The experiments are done under high medium and lowspeed manoeuvre [26] The same input steering wheel anglebetween simulation and experiment is given Therefore theresult obtained from real time control is compared with theelectronic power steering (EPS) system [15]
41 Motor Model Validation for Steering Wheel SystemAlthough the model of steering wheel system has been
12 International Journal of Vehicular Technology
Stee
ring
whe
el sy
stem
DC motor
Current
Encoder
16 bit card PCI 6221
Analogoutput
Analoginput
Counterchannel
TargetPC
Monitortarget
PC
Crossover cable
Host PC
Front axle systemand
linear vehicle model
Hardware part Simulation part
Figure 15 Block diagram of the steering wheel-XPC target interfaced
Steering wheel model
120575m1
120575m1
ia1
r(t)
Figure 16 Block diagrammotormodel validation for steeringwheelsystem
derived the model validation is necessary to ensure thesystem is fairly able to adequate for real time applicationThis section describes the motor model validation of steeringwheel system The block diagram is shown in Figure 16
The input signal is 119903(119905) and the output is rate change ofmotor ( 120575
1198981) motor position (1205751198981) and the motor current
( 1198941198981) The measured output response is then used and com-
pared with simulation to estimate the parameter of the steer-ing wheel model using the Matlab tool software The mea-sured output data are taken based on the open loop systemwhen the input to the system is sinusoidal signal withamplitude of plusmn45∘ for 10 sec The Gaussian method is usedto estimate the parameters of the steering wheel motorThe comparison between simulation and measured outputis shown in Figure 17 It is obtained that the output of thesimulation is almost identical to the measured output How-ever small difference errors occur due to the small parametricuncertainties that have not been modelled The parameterestimation value is illustrated in Table 6
42 Noise Filtering Both analog and digital devices have atrait of noise or unwanted features that have effect on thesystem performance such as electromagnetic interferenceThe noise could be either random or white noise Attenuationof this noise is often as a primary goal in control systemapplication and in particular when controlling the processsystem The average value of noise typically is zero whichwill give misreading information to the process and it wouldnot be possible to control the process In order to reduce thenoise inference filtering technique is used A filter is used toremove unwanted noise It removes some noise frequenciesto suppress interfacing signal and reduce background noise
Table 6 Parameter estimation for DCmotor steering wheel system
Parameter Value Estimation value Unit1198771 4405 564 ohm1198711 0010 0017 henry119870119904119898
1459 0024 Nm1198691198981 0002 00036 (kgm2)1198871198981 0004 00068 Nm(rads)1198961199041 165 0025 Nm
There are many types of filtering techniques such as low passfilter that has been used in this study The low pass filter isused to filter out the interference noise from the output of thecurrent sensor
Figure 18 shows the comparison of the output responsefor a current sensor with and without using filters Withouta filter the current sensor has high amplitude interferencenoise However with low pass filter under the frequency of5Hz the interference noise is reduced and the fluctuations areremoved Thus this provides a smoother output response
43 Effect on Compensation Torque A steering torque com-parison without and with compensation torque is shown inFigure 19The results for without compensation torque showsthat the steering torque response curve is far away fromEPS steering system with different amplitude as shown inFigure 19(a) This happens due to less of derivative reactiontorque from the steering wheel motor Figure 19(b) showssteering torque response with compensation torque
It is observed that by adding the additional compensationtorque the steering torque response curve trends fairly closeto EPS steering system Thus it helps to improve the driversteering feel during manoeuvre and stabilize the systemHowever with the wide range adjustment gains of the com-pensation torque it could improve the steering torqueresponse
44 Medium Speed Manoeuvre The experimental steeringtorque results at medium speed manoeuvre are shown inFigure 20(b) when the steering wheel angle is plusmn100∘ as shown
International Journal of Vehicular Technology 13
0 2 4 6 8 10minus50
0
50
Time (s)
Rate
chan
ge m
otor
pos
ition
(deg
s)
MeasurementSimulation
(a) Rate change of motor position
0 2 4 6 8 100
10
20
30
40
50
60
Time (s)
Mot
or p
ositi
on (d
eg)
Measurement
Simulation
(b) Position of the motor
0 2 4 6 8 10minus01
minus005
0
005
01
015
Mot
or cu
rren
t (A
)
MeasurementSimulation
Time (s)
(c) Current of the motor
Figure 17 Motor model validation comparison between simulation and measurement
With filter Without filter
Time (s)
Curr
ent (
A)
008
006
004
002
0
minus002
minus004
minus006
minus0081 15 2 25 3 35 4 45 5 55
Figure 18 Noise filtering using low pass filter
in Figure 20(a) From the results the experimental steeringtorque response is almost identical to the EPS steering systemHowever it has small different torque amplitudes These
occur due to the fact that mechanism or friction from themechanical gearbox motor is not modelled in proposedcontrol algorithm
14 International Journal of Vehicular Technology
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
minus152 3 4 5 6 7 8 9 10
(a) Without compensation torque
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
21 3 4 5 6 7 8 9 10
(b) With compensation torque
Figure 19 Steering torque influence on compensation torque
minus80
minus60
minus40
minus20
0
20
40
60
80
100
Stee
ring
whe
el an
gle (
deg)
Time (s)3 4 5 6 7 8 9 10
(a) Steering wheel angle at 80 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
15
20
5
0
minus5
minus10
minus15
minus203 4 5 6 7 8 9 10
(b) Steering torque estimation at 80 kmh
Figure 20 Steering torque estimation and steering wheel angle response at 80 kmh
45 High Speed Manoeuvre At high speed manoeuvre theinput steering wheel angle that is more than plusmn45∘ is very dan-gerous [12] Therefore in this experiment the input steeringwheel angle is less than plusmn45∘ as shown in Figure 21(a) Basedon the results shown in Figures 21(b) and 21(c) the experi-mental steering torque is obtained fairly similar to the EPSsteering system Although the steering wheel angle is at smallangles the influence of vehicle speed cause the steering torqueis increased Moreover at higher speed the vehicle has moreunder steer propensity [9] This increased the self aligningtorque in proportional high torque at front axle motorThusthe driver will sense stiffer on the steering wheel duringmanoeuvre
In general the centering of the steering wheel perfor-mance or steering wheel returnability is better when thesteering torque gradient is large and the hysteresis is small [9]Based on this fact the result is shown in Figure 21(c) whereby
the hysteresis of the steering wheel returnability based onexperimental result is smaller compared to EPS steeringsystem In addition the steering torque gradient is larger thanEPS steering torque Therefore it is considered that the pro-posed feedback torque algorithmnot only achieved a steeringfeel but also improved the steering wheel returnability
46 Lower Speed Manoeuvre The steering torque results atlow speed manoeuvre are shown in Figures 22(b) and 22(c)when the input steering wheel is as shown in Figure 22(a)The vehicle cruised at constant speed of 10 kmh
The experimental steering torque is almost comparableto EPS torque but the amplitude is decreases The torque isdecreased due to the fact that the vehicle speed is reducedwhich is effect on reducing the demand to the total feedbacktorque Moreover the large turning of the steering wheelincreases the backlash of the motor which then contributes
International Journal of Vehicular Technology 15
2 3 4 5 6 7 8 9 10minus30
minus20
minus10
0
10
20
30
40
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle at 120 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steeringSimulationExperiment
10
5
0
minus5
minus10
3 4 5 6 7 8 9
(b) Steering torque estimation at 120 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment10
15
5
0
minus5
minus10
minus15minus20 minus10 0 10 20 30 40
Steering wheel angle (deg)
(c) Steering torque estimation against steering wheel angle at 120 kmh
Figure 21 Steering torque estimation response at 120 kmh
to the decrease of the motor torque Thus driver will sensesofter during steering the wheel during maneuver Howeverit has small different torque against EPS steering Wide rangeadjustment of gains in proposed control algorithm may helpimprove the steering torque response
3 4 5 6 7 8 9 10minus150
minus100
minus50
0
50
100
150
200
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
10
15
5
0
minus5
minus10
minus15
Time (s)2 3 4 5 6 7 8 9 10
(b) Steering torque estimation at 10 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
Steering wheel angle (deg)
10
8
6
4
2
0
minus2
minus4
minus6
minus8
minus10
minus100 minus50 0 50 100 150
(c) Steering torque estimation against steering wheel angle at 10 kmh
Figure 22 Steering torque estimation response at 10 kmh
5 Conclusion and Future Work
The proposed force feedback torque estimation and controlalgorithm have been described The proposed control algo-rithm is considered as an effort to generate the feedback
16 International Journal of Vehicular Technology
torque for driver steering feel in the vehicle SBW systemThe mathematical model of SBW subsystem consists ofsteering wheel and front axle systems that are modelledThe effectiveness of the control algorithm is verified throughthe experimental hardware in the loop (HIL) interfacedwith Matlab XPC target toolbox software According to theexperimental results the proposed control algorithm is ableto generate the feedback torque and the driver is able to sensesteering feel similar to that obtained in EPS steering systemThe LQR controller with gain scheduling based on steeringwheel angle and vehicle speed function provides better torquecontrol that allows rejection of the uncertainty torque froma road condition Furthermore the steering feel gain is ableto increase and lower the steering torque with improvedsteering stability at high speed and it improved the vehiclemanoeuvre at low speed It is also found that compensationtorque is able to improve the feedback torque and to stabilizethe system However it has small different torque magnitudethat occurs due to the backlash of steering wheel motor anduncertainty disturbance of mechanical properties Howeverwith a wide range adjustment of the gains on the proposedcontrol algorithm it may improve and provide realisticsteering torque for better driver steering feel
In future work the proposed control algorithm will betested in real vehicle in order to have more realistic steeringtorque for further investigation and analysis
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
Acknowledgments
This work is fully supported by the Ministry of Higher Edu-cation Malaysia and University Technology Malaysia underFRGS (vote 4F370) and Research University Grant (vote00G64) This work is also supported by Proton HoldingBerhad
References
[1] Y Yao ldquoVehicle steer-by-wire system controlrdquo SAE TechnicalPaper 2006-01-1175 2006
[2] E A Kumar D Dinesh and N Kamble ldquoAn overview of activefront steering systemrdquo International Journal of Scientific ampEngineering Research vol 3 no 6 2012
[3] S M H Fahami H Zamzuri S A Mazlan and N B Zulkar-nain ldquoThe design of vehicle active front steering based on steerby wire systemrdquo Advanced Science Letters vol 19 no 1 pp 61ndash65 2013
[4] S M H Fahami H Zamzuri S A Mazlan and S A SaruchildquoThe variable steering ratio for vehicle steer by wire systemusing hyperbolic tangent methodrdquo Applied Mechanics andMaterials vol 575 pp 781ndash784 2014
[5] T Kaufmann S Milsap B Murray and J Petrowski ldquoDevelop-ment experience with steer by wirerdquo in Proceedings of the SAEInternational Congress and Exhibition SAE 2001-01-2479 2001
[6] S Amberkare F Bolourchi J Demerly and S Millsap ldquoAcontrol system methodology for steer by wire systemsrdquo SAETechnical Paper 2004-01-1106 2004
[7] A Baviskar J R Wagner D M Dawson D Braganza andP Setlur ldquoAn adjustable steer-by-wire haptic-interface trackingcontroller for ground vehiclesrdquo IEEE Transactions on VehicularTechnology vol 58 no 2 pp 546ndash554 2009
[8] M Segawa S Kimura T Kada and S Nakona ldquoA study of reac-tive torque control for steer bywire systemrdquo inProceedings of theInternational Symposium on Advanced Vehicle Control (AVECrsquo02) pp 653ndash658 Hiroshima Japan September 2002
[9] S W Oh H C Chae S C Yun and C S Han ldquoThe design ofa controller for the steer-by-wire systemrdquo JSME InternationalJournal Series C Mechanical Systems Machine Elements andManufacturing vol 47 no 3 pp 896ndash907 2004
[10] C J Kim J H Jang S K Oh J Y Lee and J K HedrickldquoDevelopment of a control algorithm for a rack-actuating steer-by-wire system using road information feedbackrdquo Proceedingsof the Institution of Mechanical Engineers Part D Journal ofAutomobile Engineering vol 222 no 9 pp 1559ndash1571 2008
[11] S Asai H Kuroyanagi and S Takeuchi ldquoDevelopment of asteer-by-wire system with force feedback using a disturbanceobserverrdquo inProceedings of the SAEWorldCongressampExhibitionon Steering amp Suspension Technology Detroit Mich USA 2004
[12] T J Park C S Han and S H Lee ldquoDevelopment of the elec-tronic control unit for the rack-actuating steer-by-wire usingthe hardware-in-the-loop simulation systemrdquo Mechatronicsvol 15 no 8 pp 899ndash918 2005
[13] D Odenhtal T Bunte H D Heitzer and C Eicker ldquoHow tomake steer-by-wire feel like power steeringrdquo in Proceedings ofthe 15th IFACWorld Congress Barcelona Spain July 2000
[14] H Na C Zong and D Hu ldquoInvestigations on corneringcontrol algorithm design and road feeling optimization for asteer-by-wire vehiclerdquo in Proceedings of the IEEE InternationalConference on Mechatronics and Automation (ICMA rsquo09) pp3246ndash3251 Changchun China August 2009
[15] S Ancha A Baviskar J RWagner andDMDawson ldquoGroundvehicle steering systems modelling control and analysis ofhydraulic electric and steer-by-wire configurationsrdquo Interna-tional Journal of Vehicle Design vol 44 no 1-2 pp 188ndash2082007
[16] P Yih and J C Gerdes ldquoModification of vehicle handlingcharacteristics via steer-by-wirerdquo IEEE Transactions on ControlSystems Technology vol 13 no 6 pp 965ndash976 2005
[17] S M H Fahami H Zamzuri S A Mazlan and M A ZakarialdquoModeling and simulation of vehicle steer by wire systemrdquoin Proceedings of the IEEE Symposium on Humanities Scienceand Engineering Research (SHUSER rsquo12) pp 765ndash770 KualaLumpur Malaysia June 2012
[18] S Anwar and L Chen ldquoAn analytical redundancy-based faultdetection and isolation algorithm for a road-wheel controlsubsystem in a steer-by-wire systemrdquo IEEE Transactions onVehicular Technology vol 56 no 5 pp 2859ndash2869 2007
[19] M T Do Z Man C Zhang H Wang and F S Tay ldquoRobustsliding mode-based learning control for steer-by-wire systemsin modern vehiclesrdquo IEEE Transactions on Vehicular Technol-ogy vol 63 no 2 pp 580ndash590 2014
[20] J Duan R Wang and Y Yu ldquoResearch on control strategies ofsteer-by-wire systemrdquo in Proceedings of the International Con-ference on Intelligent Computation Technology and Automation(ICICTA rsquo10) vol 2 pp 1122ndash1125 May 2010
International Journal of Vehicular Technology 17
[21] B Zheng and S Anwar ldquoYaw stability control of a steer-by-wireequipped vehicle via active front wheel steeringrdquoMechatronicsvol 19 no 6 pp 799ndash804 2009
[22] S Sahoo S C Subramanian and S Srivastava ldquoDesign andimplementation of a controller for navigating an autonomousground vehiclerdquo in Proceedings of the 2nd International Confer-ence on Power Control and Embedded Systems (ICPCES rsquo12) pp1ndash6 IEEE Allahabad India December 2012
[23] T D Gillespie Fundamentals of Vehicle Dynamics Society ofAutomotive Engineers 1992
[24] M H Lee S K Ha J Y Choi and K S Yoon ldquoImprovement ofthe steering feel of an electric power steering system by torquemap modificationrdquo Journal of Mechanical Science and Technol-ogy vol 19 no 3 pp 792ndash801 2005
[25] R Pastorino M A Naya J A Perez and J Cuadrado ldquoGearedPM coreless motor modelling for driverrsquos force feedback insteer-by-wire systemsrdquo Mechatronics vol 21 no 6 pp 1043ndash1054 2011
[26] P Koehn andM Eckrich ldquoActive steeringmdashthe BMWapproachtowards modern steering technologyrdquo SAE Technical PaperSeries 2004-01-1105 SAE International 2004
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
8 International Journal of Vehicular Technology
0 2 4 6 8 10minus6
minus4
minus2
0
2
4
6
Time (s)
Torq
ue (N
m)
Inertia torque Damping torque
Figure 8 Inertia and damping torque
0 2 4 6 8 10Time (s)
Forc
e fee
dbac
k to
rque
(Nm
)
15
10
5
0
minus5
minus10
minus15
minus20
minus25
10kmh
50kmh100 kmh120 kmh
135kmh
Figure 9 Proposed force feedback torque response at differentvehicle speed
torque response based on (24) The torque is increasedaccording to the vehicle speed The higher vehicle speed willincrease the feedback torque and vice versa due to the samefact that is proposed by torque map
31 Linear Quadratic Regulator (LQR) Controller The LQRcontroller provides the best possible performance to controland stabilize the system by changing the location of the polesof the system to the optimal location for a time responseovershoot and steady stateThis is done by designing the statefeedback control119870 thatminimized cost function 119869 to stabilizethe system In this paper the LQR controller is used firstlyto control the front axle motor angle for the wheel synchro-nization and secondly to control the steering wheel motorfor a feedback torque
The wheel synchronization or directional control is abasic characteristic for steering function where a front tire
+
minus
sum
k1
k2
k3
x1
x2
x3
Plantr(t)
LQR controller
Figure 10 Basic control block diagram of LQR controller
angle should follow driver input steering wheel angle accord-ing to the steering ratio This is done by controlling theangle of the front axle motor In the LQR control schemesa feedback gain matrix is designed to achieve some compro-mise between the use of control effort the magnitude andthe speed of response that will guarantee a stable systemFigure 10 shows the control block diagram of the LQRcontroller
The optimal LQRproblem is often definedmore generallyand consists of finding controller transfer matrix that mini-mizes the cost function 119869 that is written as follows
119869 = int
prop
0(119909119879
119876119909+119906119879
119877119906) 119889119905 (31)
where 119909119879 and119880119879 respectively represent the transpose of the
state and input vector The 119876 and 119877 are positive definite (orpositive-semi definite) real symmetric matrix The 119876 and 119877
are usually chosen to be diagonal matrix written as follows
119876119894119894=
1maximum acceptable value of 1199092
119894
119877119895119895
=1
maximum acceptable value of 1199062119895
(32)
where
119894 = 120576 1 2 ℓ
119895 = 120576 1 2 119895 (33)
In essence Brysonrsquos rule scales the variables that appear in 119869 sothat the maximum acceptable value for each term is oneThisis especially important when the units used for the differentcomponents of 119906 and 119909 make the values for these variablesnumerically very different from each other
Assume that the steering ratio is 15 1 and the gains ofthe LQR controller are [01 098 0008] Figure 11 shows aresponse of the front axle system when the input steeringwheel angle is a step function as shown in Figure 11(a)
The results show that the angle of the front axle motorhas a fast response compared to the EPS steering system
International Journal of Vehicular Technology 9
0 2 4 6 8 100
2
4
6
8
10
12
14
16
18
20
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle
0 2 4 6 8 100
2
4
6
8
10
12
14
16
18
Time (s)
Mot
orc
olum
n an
gle (
deg)
EPS steeringmdashcolumn shaft angle
SBW systemmdashmotor angle
(b) Front axle motor angle
0 2 4 6 8 10
Rack
disp
lace
men
t (m
)
SBW system
EPS steering
Time (s)
18
16
14
12
10
8
6
4
2
0
times10minus3
(c) Rack displacement
0 2 4 6 8 100
02
04
06
08
1
12
14Fr
ont t
ire an
gle (
deg)
SBW system
EPS steering
Time (s)
(d) Front tire angle
Figure 11 Front axle system response for wheel synchronization
Moreover the LQR controller is able to reduce the delaycaused by the self aligning torque created from the contactbetween the tire and road surfaces This not only improvedthe vehicle stability but also ensured that the driver has amore confident level during manoeuvre Compared to theEPS steering system the delay caused by mechanical columnshaft connected to the steering wheel system Furthermore itcan be seen that the front tire angle is at almost fairly close to1∘ as shown in Figure 11(d) However it has small chatteringdue to the self aligning torque However with wide rangeadjustment gains of the LQR controller the steering responsecould be improved
32 Gain Scheduling (GS) with LQR Controller In certainsituations the dynamic behaviour of the process changed inaccordance with the conditions of the process This situation
could affect the stability of the system if not to control prop-erly Thus it is possible to change the gains of the controllerby monitoring the operating conditions of the process tohave a stable system The gain scheduling (GS) techniqueis used to change the gain of the LQR controller in order tocontrol the feedback torque and the steering feel gain (119896feel)is used to increase and decrease the torque It is important toincrease the feedback torque at high speed for steering stabil-ity and to decrease it in low speed especially during parkingto improve vehicle manoeuvre [9] This is based on the steer-ingwheel angle and vehicle speed function in order to achievethe feedback torque for the real driver steering feel
The relationship between a gain scheduling with LQRcontroller is shown in Table 4 The vehicle speed is dividedinto 3 categories at low speed (0ndash30 kmh) medium speed(31ndash100 kmh) and high speed (101ndash120 kmh) The maxi-mum vehicle speed is limited to 120 kmhThe steering wheel
10 International Journal of Vehicular Technology
0 2 4 6 8 10minus50
minus40
minus30
minus20
minus10
0
10
20
30
40
50
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Input steering wheel angle
0 1 2 3 4 5 6 7 8 9 10minus15
minus10
minus5
0
5
10
Time (s)
Stee
ring
torq
ue (N
m) LQR
LQR + GS
Reference
(b) 120583 = 1
0 1 2 3 4 5 6 7 8 9 10
minus15
minus10
minus5
0
5
10
15
Stee
ring
torq
ue (N
m) LQR
Reference
Time (s)
LQR + GS
(c) 120583 = 14
0 1 2 3 4 5 6 7 8 9 10minus20
minus15
minus10
minus5
0
5
10
15
20
Stee
ring
torq
ue (N
m) LQR LQR + GS
Reference
Time (s)
(d) 120583 = 17
Figure 12 Comparison of torque response of LQR and LQR + GS controller at different 120583
Table 4 The gains of LQR controller and steering feel
119881 (kmh) 120575119904119908
(degree) 1198961 1198962 1198963 119896feel
(0ndash30) (plusmn180∘ lt 120575119904119908
lt plusmn405∘) 003 065 0001 014(31ndash100) (plusmn46∘ lt 120575
119904119908lt plusmn179∘) 003 084 0001 060
(101ndash120) 120575119904119908
le +45∘ or 120575119904119908
le minus45∘ 005 095 0001 095
angle is also divided into 3 categories according to vehiclespeed condition The notation ldquo+rdquo means that the driverturned the steering wheel clockwise and notation ldquominusrdquo is usedfor counter clockwise in which the maximum steering wheelturn is plusmn405∘
The gains of the LQR controller are defined using Brysonrsquosrule method and the steering feel gain is defined usingintuitive and has a range from 01 to 1 This range has beendefined for various speeds 10 kmh 80 kmh and 120 kmhafter comparing with EPS steering system [15] For exampleif the driver turns a steering wheel between (plusmn180∘ lt 120575
119904119908lt
plusmn405∘) and vehicle speed is 10 kmh the total feedback torque
is multiplied by steering feel gain 119896feel = 014 Thus thefeedback torque is decreased which results in making thedriver feel it easier turning the steering wheel and improvesvehicle manoeuvre The feedback torque is regulated by LQRcontroller when the gains are 1198961 = 003 1198962 = 065 and1198963 = 0001 The gains are the optimum that has been assuredby providing a better torque control based on vehicle speedand steering wheel angle function
Comparing torque response between LQR and LQR +GScontroller for different coefficient (120583) is shown in Figure 12when the input steering wheel angle is a double lane changemanoeuvre as shown in Figure 12(a) The vehicle speed isset to be 80 kmh Based on results obtained the LQR + GStechnique offers an improvement 90 and the curve trend isalmost similar to reference torque compared to the LQR con-troller This is because the gains of LQR controller changedbased on the steering wheel angle and vehicle speed param-eter which is able to reject the uncertainty torque of a roadcondition Meanwhile with the fixed gains of the LQR con-troller it has an excessive torque that can cause the system tobe unstable The RMS torque values are illustrated in Table 5
International Journal of Vehicular Technology 11
Table 5 Comparison of RMS torque values of LQR and LQR + GScontroller
Reference LQR LQR + GS120583 = 1 56361 50795 55450120583 = 14 61218 68339 57243120583 = 17 66184 72888 62080
minus50 0 50minus15
minus10
minus5
0
5
10
15
Steering angle (deg)
Stee
ring
torq
ue (N
m) Feel gain = 5
Feel gain = 01Feel gain = 05
Feel gain = 1
(a) Different steering feel gain at 80 kmh
minus50 minus40 minus30 minus20 minus10 0 10 20 30 40 50
minus15
minus10
minus5
0
5
10
15
Steering wheel angle (deg)
Stee
ring
torq
ue (N
m)
80kmh kfeel = 060
120 kmh kfeel = 095
10kmh kfeel = 014
(b) Different steering feel gain at various speed
Figure 13 Steering feel gains response at various vehicle speed
Figure 13(a) shows the steering torque with respect tosteering wheel angle when different steering feel gain is usedat various vehicle speeds It is observed that when the steeringfeel gain is increased at maximum range which is 1 thesteering torque will increase and driver will sense stiffer onthe steering wheel at high speed manoeuvre Meanwhilewhen decreasing the steering feel gain this will result inreducing the steering torqueThus the driver will sense softer
Target monitor
Host PC
Cross over cableConnector block
PCI 6221
Power supplyCurrent sensor
DC motor driver
Figure 14 Steering wheel system hardware overview
during turning the steering wheel at the low speed manoeu-vre However when the steering feel gain is 5 which exceedsits maximum limit the steering torque becomes unstableFigure 13(b) shows the steering feel gains at various speedsAt high speed of 120 kmh the steering feel gain is increasedand this will increase the steering torque for steering stabilityMeanwhile at low speed the steering feel gain is decreasedwhich will help improve vehicle manoeuvre especially duringparking Based on the proposed gain scheduling with theLQR controller not only does it provide better control on thesteering torque but also it is optional for a designer to have awide range adjustment to vary the steering torque for a betterdriver steering feel and to improve vehicle manoeuvre
4 Hardware in the Loop (HIL) and Result
For hardware in the loop (HIL) setup the front axle systemand linear vehicle model are simulated on the host computerThe steering wheel system is replaced by HIL mechanismThe position of the steering wheel motor is measured usinga rotary encoder with 500 pulses per revolution To measurethe steering wheel torque the current sensor is connectedin series with the DC motor driver The steering wheel DCmotor has a torque 129Nm with 100 rpm The single tracklinear vehicle model is used to generate the estimation selfaligning torque as element to the feedback torque and distur-bance to the front axle system The HIL system is executedunder 1ms sampling time and interfaced using Matlab-XPCtarget toolbox configuration as shown in Figures 14 and 15
The assumptions through the experiment are listed asfollows
(i) the vehicle cruises at constant speed(ii) the condition of dry asphalt road (120583 = 085)
The experiments are done under high medium and lowspeed manoeuvre [26] The same input steering wheel anglebetween simulation and experiment is given Therefore theresult obtained from real time control is compared with theelectronic power steering (EPS) system [15]
41 Motor Model Validation for Steering Wheel SystemAlthough the model of steering wheel system has been
12 International Journal of Vehicular Technology
Stee
ring
whe
el sy
stem
DC motor
Current
Encoder
16 bit card PCI 6221
Analogoutput
Analoginput
Counterchannel
TargetPC
Monitortarget
PC
Crossover cable
Host PC
Front axle systemand
linear vehicle model
Hardware part Simulation part
Figure 15 Block diagram of the steering wheel-XPC target interfaced
Steering wheel model
120575m1
120575m1
ia1
r(t)
Figure 16 Block diagrammotormodel validation for steeringwheelsystem
derived the model validation is necessary to ensure thesystem is fairly able to adequate for real time applicationThis section describes the motor model validation of steeringwheel system The block diagram is shown in Figure 16
The input signal is 119903(119905) and the output is rate change ofmotor ( 120575
1198981) motor position (1205751198981) and the motor current
( 1198941198981) The measured output response is then used and com-
pared with simulation to estimate the parameter of the steer-ing wheel model using the Matlab tool software The mea-sured output data are taken based on the open loop systemwhen the input to the system is sinusoidal signal withamplitude of plusmn45∘ for 10 sec The Gaussian method is usedto estimate the parameters of the steering wheel motorThe comparison between simulation and measured outputis shown in Figure 17 It is obtained that the output of thesimulation is almost identical to the measured output How-ever small difference errors occur due to the small parametricuncertainties that have not been modelled The parameterestimation value is illustrated in Table 6
42 Noise Filtering Both analog and digital devices have atrait of noise or unwanted features that have effect on thesystem performance such as electromagnetic interferenceThe noise could be either random or white noise Attenuationof this noise is often as a primary goal in control systemapplication and in particular when controlling the processsystem The average value of noise typically is zero whichwill give misreading information to the process and it wouldnot be possible to control the process In order to reduce thenoise inference filtering technique is used A filter is used toremove unwanted noise It removes some noise frequenciesto suppress interfacing signal and reduce background noise
Table 6 Parameter estimation for DCmotor steering wheel system
Parameter Value Estimation value Unit1198771 4405 564 ohm1198711 0010 0017 henry119870119904119898
1459 0024 Nm1198691198981 0002 00036 (kgm2)1198871198981 0004 00068 Nm(rads)1198961199041 165 0025 Nm
There are many types of filtering techniques such as low passfilter that has been used in this study The low pass filter isused to filter out the interference noise from the output of thecurrent sensor
Figure 18 shows the comparison of the output responsefor a current sensor with and without using filters Withouta filter the current sensor has high amplitude interferencenoise However with low pass filter under the frequency of5Hz the interference noise is reduced and the fluctuations areremoved Thus this provides a smoother output response
43 Effect on Compensation Torque A steering torque com-parison without and with compensation torque is shown inFigure 19The results for without compensation torque showsthat the steering torque response curve is far away fromEPS steering system with different amplitude as shown inFigure 19(a) This happens due to less of derivative reactiontorque from the steering wheel motor Figure 19(b) showssteering torque response with compensation torque
It is observed that by adding the additional compensationtorque the steering torque response curve trends fairly closeto EPS steering system Thus it helps to improve the driversteering feel during manoeuvre and stabilize the systemHowever with the wide range adjustment gains of the com-pensation torque it could improve the steering torqueresponse
44 Medium Speed Manoeuvre The experimental steeringtorque results at medium speed manoeuvre are shown inFigure 20(b) when the steering wheel angle is plusmn100∘ as shown
International Journal of Vehicular Technology 13
0 2 4 6 8 10minus50
0
50
Time (s)
Rate
chan
ge m
otor
pos
ition
(deg
s)
MeasurementSimulation
(a) Rate change of motor position
0 2 4 6 8 100
10
20
30
40
50
60
Time (s)
Mot
or p
ositi
on (d
eg)
Measurement
Simulation
(b) Position of the motor
0 2 4 6 8 10minus01
minus005
0
005
01
015
Mot
or cu
rren
t (A
)
MeasurementSimulation
Time (s)
(c) Current of the motor
Figure 17 Motor model validation comparison between simulation and measurement
With filter Without filter
Time (s)
Curr
ent (
A)
008
006
004
002
0
minus002
minus004
minus006
minus0081 15 2 25 3 35 4 45 5 55
Figure 18 Noise filtering using low pass filter
in Figure 20(a) From the results the experimental steeringtorque response is almost identical to the EPS steering systemHowever it has small different torque amplitudes These
occur due to the fact that mechanism or friction from themechanical gearbox motor is not modelled in proposedcontrol algorithm
14 International Journal of Vehicular Technology
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
minus152 3 4 5 6 7 8 9 10
(a) Without compensation torque
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
21 3 4 5 6 7 8 9 10
(b) With compensation torque
Figure 19 Steering torque influence on compensation torque
minus80
minus60
minus40
minus20
0
20
40
60
80
100
Stee
ring
whe
el an
gle (
deg)
Time (s)3 4 5 6 7 8 9 10
(a) Steering wheel angle at 80 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
15
20
5
0
minus5
minus10
minus15
minus203 4 5 6 7 8 9 10
(b) Steering torque estimation at 80 kmh
Figure 20 Steering torque estimation and steering wheel angle response at 80 kmh
45 High Speed Manoeuvre At high speed manoeuvre theinput steering wheel angle that is more than plusmn45∘ is very dan-gerous [12] Therefore in this experiment the input steeringwheel angle is less than plusmn45∘ as shown in Figure 21(a) Basedon the results shown in Figures 21(b) and 21(c) the experi-mental steering torque is obtained fairly similar to the EPSsteering system Although the steering wheel angle is at smallangles the influence of vehicle speed cause the steering torqueis increased Moreover at higher speed the vehicle has moreunder steer propensity [9] This increased the self aligningtorque in proportional high torque at front axle motorThusthe driver will sense stiffer on the steering wheel duringmanoeuvre
In general the centering of the steering wheel perfor-mance or steering wheel returnability is better when thesteering torque gradient is large and the hysteresis is small [9]Based on this fact the result is shown in Figure 21(c) whereby
the hysteresis of the steering wheel returnability based onexperimental result is smaller compared to EPS steeringsystem In addition the steering torque gradient is larger thanEPS steering torque Therefore it is considered that the pro-posed feedback torque algorithmnot only achieved a steeringfeel but also improved the steering wheel returnability
46 Lower Speed Manoeuvre The steering torque results atlow speed manoeuvre are shown in Figures 22(b) and 22(c)when the input steering wheel is as shown in Figure 22(a)The vehicle cruised at constant speed of 10 kmh
The experimental steering torque is almost comparableto EPS torque but the amplitude is decreases The torque isdecreased due to the fact that the vehicle speed is reducedwhich is effect on reducing the demand to the total feedbacktorque Moreover the large turning of the steering wheelincreases the backlash of the motor which then contributes
International Journal of Vehicular Technology 15
2 3 4 5 6 7 8 9 10minus30
minus20
minus10
0
10
20
30
40
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle at 120 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steeringSimulationExperiment
10
5
0
minus5
minus10
3 4 5 6 7 8 9
(b) Steering torque estimation at 120 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment10
15
5
0
minus5
minus10
minus15minus20 minus10 0 10 20 30 40
Steering wheel angle (deg)
(c) Steering torque estimation against steering wheel angle at 120 kmh
Figure 21 Steering torque estimation response at 120 kmh
to the decrease of the motor torque Thus driver will sensesofter during steering the wheel during maneuver Howeverit has small different torque against EPS steering Wide rangeadjustment of gains in proposed control algorithm may helpimprove the steering torque response
3 4 5 6 7 8 9 10minus150
minus100
minus50
0
50
100
150
200
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
10
15
5
0
minus5
minus10
minus15
Time (s)2 3 4 5 6 7 8 9 10
(b) Steering torque estimation at 10 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
Steering wheel angle (deg)
10
8
6
4
2
0
minus2
minus4
minus6
minus8
minus10
minus100 minus50 0 50 100 150
(c) Steering torque estimation against steering wheel angle at 10 kmh
Figure 22 Steering torque estimation response at 10 kmh
5 Conclusion and Future Work
The proposed force feedback torque estimation and controlalgorithm have been described The proposed control algo-rithm is considered as an effort to generate the feedback
16 International Journal of Vehicular Technology
torque for driver steering feel in the vehicle SBW systemThe mathematical model of SBW subsystem consists ofsteering wheel and front axle systems that are modelledThe effectiveness of the control algorithm is verified throughthe experimental hardware in the loop (HIL) interfacedwith Matlab XPC target toolbox software According to theexperimental results the proposed control algorithm is ableto generate the feedback torque and the driver is able to sensesteering feel similar to that obtained in EPS steering systemThe LQR controller with gain scheduling based on steeringwheel angle and vehicle speed function provides better torquecontrol that allows rejection of the uncertainty torque froma road condition Furthermore the steering feel gain is ableto increase and lower the steering torque with improvedsteering stability at high speed and it improved the vehiclemanoeuvre at low speed It is also found that compensationtorque is able to improve the feedback torque and to stabilizethe system However it has small different torque magnitudethat occurs due to the backlash of steering wheel motor anduncertainty disturbance of mechanical properties Howeverwith a wide range adjustment of the gains on the proposedcontrol algorithm it may improve and provide realisticsteering torque for better driver steering feel
In future work the proposed control algorithm will betested in real vehicle in order to have more realistic steeringtorque for further investigation and analysis
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
Acknowledgments
This work is fully supported by the Ministry of Higher Edu-cation Malaysia and University Technology Malaysia underFRGS (vote 4F370) and Research University Grant (vote00G64) This work is also supported by Proton HoldingBerhad
References
[1] Y Yao ldquoVehicle steer-by-wire system controlrdquo SAE TechnicalPaper 2006-01-1175 2006
[2] E A Kumar D Dinesh and N Kamble ldquoAn overview of activefront steering systemrdquo International Journal of Scientific ampEngineering Research vol 3 no 6 2012
[3] S M H Fahami H Zamzuri S A Mazlan and N B Zulkar-nain ldquoThe design of vehicle active front steering based on steerby wire systemrdquo Advanced Science Letters vol 19 no 1 pp 61ndash65 2013
[4] S M H Fahami H Zamzuri S A Mazlan and S A SaruchildquoThe variable steering ratio for vehicle steer by wire systemusing hyperbolic tangent methodrdquo Applied Mechanics andMaterials vol 575 pp 781ndash784 2014
[5] T Kaufmann S Milsap B Murray and J Petrowski ldquoDevelop-ment experience with steer by wirerdquo in Proceedings of the SAEInternational Congress and Exhibition SAE 2001-01-2479 2001
[6] S Amberkare F Bolourchi J Demerly and S Millsap ldquoAcontrol system methodology for steer by wire systemsrdquo SAETechnical Paper 2004-01-1106 2004
[7] A Baviskar J R Wagner D M Dawson D Braganza andP Setlur ldquoAn adjustable steer-by-wire haptic-interface trackingcontroller for ground vehiclesrdquo IEEE Transactions on VehicularTechnology vol 58 no 2 pp 546ndash554 2009
[8] M Segawa S Kimura T Kada and S Nakona ldquoA study of reac-tive torque control for steer bywire systemrdquo inProceedings of theInternational Symposium on Advanced Vehicle Control (AVECrsquo02) pp 653ndash658 Hiroshima Japan September 2002
[9] S W Oh H C Chae S C Yun and C S Han ldquoThe design ofa controller for the steer-by-wire systemrdquo JSME InternationalJournal Series C Mechanical Systems Machine Elements andManufacturing vol 47 no 3 pp 896ndash907 2004
[10] C J Kim J H Jang S K Oh J Y Lee and J K HedrickldquoDevelopment of a control algorithm for a rack-actuating steer-by-wire system using road information feedbackrdquo Proceedingsof the Institution of Mechanical Engineers Part D Journal ofAutomobile Engineering vol 222 no 9 pp 1559ndash1571 2008
[11] S Asai H Kuroyanagi and S Takeuchi ldquoDevelopment of asteer-by-wire system with force feedback using a disturbanceobserverrdquo inProceedings of the SAEWorldCongressampExhibitionon Steering amp Suspension Technology Detroit Mich USA 2004
[12] T J Park C S Han and S H Lee ldquoDevelopment of the elec-tronic control unit for the rack-actuating steer-by-wire usingthe hardware-in-the-loop simulation systemrdquo Mechatronicsvol 15 no 8 pp 899ndash918 2005
[13] D Odenhtal T Bunte H D Heitzer and C Eicker ldquoHow tomake steer-by-wire feel like power steeringrdquo in Proceedings ofthe 15th IFACWorld Congress Barcelona Spain July 2000
[14] H Na C Zong and D Hu ldquoInvestigations on corneringcontrol algorithm design and road feeling optimization for asteer-by-wire vehiclerdquo in Proceedings of the IEEE InternationalConference on Mechatronics and Automation (ICMA rsquo09) pp3246ndash3251 Changchun China August 2009
[15] S Ancha A Baviskar J RWagner andDMDawson ldquoGroundvehicle steering systems modelling control and analysis ofhydraulic electric and steer-by-wire configurationsrdquo Interna-tional Journal of Vehicle Design vol 44 no 1-2 pp 188ndash2082007
[16] P Yih and J C Gerdes ldquoModification of vehicle handlingcharacteristics via steer-by-wirerdquo IEEE Transactions on ControlSystems Technology vol 13 no 6 pp 965ndash976 2005
[17] S M H Fahami H Zamzuri S A Mazlan and M A ZakarialdquoModeling and simulation of vehicle steer by wire systemrdquoin Proceedings of the IEEE Symposium on Humanities Scienceand Engineering Research (SHUSER rsquo12) pp 765ndash770 KualaLumpur Malaysia June 2012
[18] S Anwar and L Chen ldquoAn analytical redundancy-based faultdetection and isolation algorithm for a road-wheel controlsubsystem in a steer-by-wire systemrdquo IEEE Transactions onVehicular Technology vol 56 no 5 pp 2859ndash2869 2007
[19] M T Do Z Man C Zhang H Wang and F S Tay ldquoRobustsliding mode-based learning control for steer-by-wire systemsin modern vehiclesrdquo IEEE Transactions on Vehicular Technol-ogy vol 63 no 2 pp 580ndash590 2014
[20] J Duan R Wang and Y Yu ldquoResearch on control strategies ofsteer-by-wire systemrdquo in Proceedings of the International Con-ference on Intelligent Computation Technology and Automation(ICICTA rsquo10) vol 2 pp 1122ndash1125 May 2010
International Journal of Vehicular Technology 17
[21] B Zheng and S Anwar ldquoYaw stability control of a steer-by-wireequipped vehicle via active front wheel steeringrdquoMechatronicsvol 19 no 6 pp 799ndash804 2009
[22] S Sahoo S C Subramanian and S Srivastava ldquoDesign andimplementation of a controller for navigating an autonomousground vehiclerdquo in Proceedings of the 2nd International Confer-ence on Power Control and Embedded Systems (ICPCES rsquo12) pp1ndash6 IEEE Allahabad India December 2012
[23] T D Gillespie Fundamentals of Vehicle Dynamics Society ofAutomotive Engineers 1992
[24] M H Lee S K Ha J Y Choi and K S Yoon ldquoImprovement ofthe steering feel of an electric power steering system by torquemap modificationrdquo Journal of Mechanical Science and Technol-ogy vol 19 no 3 pp 792ndash801 2005
[25] R Pastorino M A Naya J A Perez and J Cuadrado ldquoGearedPM coreless motor modelling for driverrsquos force feedback insteer-by-wire systemsrdquo Mechatronics vol 21 no 6 pp 1043ndash1054 2011
[26] P Koehn andM Eckrich ldquoActive steeringmdashthe BMWapproachtowards modern steering technologyrdquo SAE Technical PaperSeries 2004-01-1105 SAE International 2004
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
International Journal of Vehicular Technology 9
0 2 4 6 8 100
2
4
6
8
10
12
14
16
18
20
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle
0 2 4 6 8 100
2
4
6
8
10
12
14
16
18
Time (s)
Mot
orc
olum
n an
gle (
deg)
EPS steeringmdashcolumn shaft angle
SBW systemmdashmotor angle
(b) Front axle motor angle
0 2 4 6 8 10
Rack
disp
lace
men
t (m
)
SBW system
EPS steering
Time (s)
18
16
14
12
10
8
6
4
2
0
times10minus3
(c) Rack displacement
0 2 4 6 8 100
02
04
06
08
1
12
14Fr
ont t
ire an
gle (
deg)
SBW system
EPS steering
Time (s)
(d) Front tire angle
Figure 11 Front axle system response for wheel synchronization
Moreover the LQR controller is able to reduce the delaycaused by the self aligning torque created from the contactbetween the tire and road surfaces This not only improvedthe vehicle stability but also ensured that the driver has amore confident level during manoeuvre Compared to theEPS steering system the delay caused by mechanical columnshaft connected to the steering wheel system Furthermore itcan be seen that the front tire angle is at almost fairly close to1∘ as shown in Figure 11(d) However it has small chatteringdue to the self aligning torque However with wide rangeadjustment gains of the LQR controller the steering responsecould be improved
32 Gain Scheduling (GS) with LQR Controller In certainsituations the dynamic behaviour of the process changed inaccordance with the conditions of the process This situation
could affect the stability of the system if not to control prop-erly Thus it is possible to change the gains of the controllerby monitoring the operating conditions of the process tohave a stable system The gain scheduling (GS) techniqueis used to change the gain of the LQR controller in order tocontrol the feedback torque and the steering feel gain (119896feel)is used to increase and decrease the torque It is important toincrease the feedback torque at high speed for steering stabil-ity and to decrease it in low speed especially during parkingto improve vehicle manoeuvre [9] This is based on the steer-ingwheel angle and vehicle speed function in order to achievethe feedback torque for the real driver steering feel
The relationship between a gain scheduling with LQRcontroller is shown in Table 4 The vehicle speed is dividedinto 3 categories at low speed (0ndash30 kmh) medium speed(31ndash100 kmh) and high speed (101ndash120 kmh) The maxi-mum vehicle speed is limited to 120 kmhThe steering wheel
10 International Journal of Vehicular Technology
0 2 4 6 8 10minus50
minus40
minus30
minus20
minus10
0
10
20
30
40
50
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Input steering wheel angle
0 1 2 3 4 5 6 7 8 9 10minus15
minus10
minus5
0
5
10
Time (s)
Stee
ring
torq
ue (N
m) LQR
LQR + GS
Reference
(b) 120583 = 1
0 1 2 3 4 5 6 7 8 9 10
minus15
minus10
minus5
0
5
10
15
Stee
ring
torq
ue (N
m) LQR
Reference
Time (s)
LQR + GS
(c) 120583 = 14
0 1 2 3 4 5 6 7 8 9 10minus20
minus15
minus10
minus5
0
5
10
15
20
Stee
ring
torq
ue (N
m) LQR LQR + GS
Reference
Time (s)
(d) 120583 = 17
Figure 12 Comparison of torque response of LQR and LQR + GS controller at different 120583
Table 4 The gains of LQR controller and steering feel
119881 (kmh) 120575119904119908
(degree) 1198961 1198962 1198963 119896feel
(0ndash30) (plusmn180∘ lt 120575119904119908
lt plusmn405∘) 003 065 0001 014(31ndash100) (plusmn46∘ lt 120575
119904119908lt plusmn179∘) 003 084 0001 060
(101ndash120) 120575119904119908
le +45∘ or 120575119904119908
le minus45∘ 005 095 0001 095
angle is also divided into 3 categories according to vehiclespeed condition The notation ldquo+rdquo means that the driverturned the steering wheel clockwise and notation ldquominusrdquo is usedfor counter clockwise in which the maximum steering wheelturn is plusmn405∘
The gains of the LQR controller are defined using Brysonrsquosrule method and the steering feel gain is defined usingintuitive and has a range from 01 to 1 This range has beendefined for various speeds 10 kmh 80 kmh and 120 kmhafter comparing with EPS steering system [15] For exampleif the driver turns a steering wheel between (plusmn180∘ lt 120575
119904119908lt
plusmn405∘) and vehicle speed is 10 kmh the total feedback torque
is multiplied by steering feel gain 119896feel = 014 Thus thefeedback torque is decreased which results in making thedriver feel it easier turning the steering wheel and improvesvehicle manoeuvre The feedback torque is regulated by LQRcontroller when the gains are 1198961 = 003 1198962 = 065 and1198963 = 0001 The gains are the optimum that has been assuredby providing a better torque control based on vehicle speedand steering wheel angle function
Comparing torque response between LQR and LQR +GScontroller for different coefficient (120583) is shown in Figure 12when the input steering wheel angle is a double lane changemanoeuvre as shown in Figure 12(a) The vehicle speed isset to be 80 kmh Based on results obtained the LQR + GStechnique offers an improvement 90 and the curve trend isalmost similar to reference torque compared to the LQR con-troller This is because the gains of LQR controller changedbased on the steering wheel angle and vehicle speed param-eter which is able to reject the uncertainty torque of a roadcondition Meanwhile with the fixed gains of the LQR con-troller it has an excessive torque that can cause the system tobe unstable The RMS torque values are illustrated in Table 5
International Journal of Vehicular Technology 11
Table 5 Comparison of RMS torque values of LQR and LQR + GScontroller
Reference LQR LQR + GS120583 = 1 56361 50795 55450120583 = 14 61218 68339 57243120583 = 17 66184 72888 62080
minus50 0 50minus15
minus10
minus5
0
5
10
15
Steering angle (deg)
Stee
ring
torq
ue (N
m) Feel gain = 5
Feel gain = 01Feel gain = 05
Feel gain = 1
(a) Different steering feel gain at 80 kmh
minus50 minus40 minus30 minus20 minus10 0 10 20 30 40 50
minus15
minus10
minus5
0
5
10
15
Steering wheel angle (deg)
Stee
ring
torq
ue (N
m)
80kmh kfeel = 060
120 kmh kfeel = 095
10kmh kfeel = 014
(b) Different steering feel gain at various speed
Figure 13 Steering feel gains response at various vehicle speed
Figure 13(a) shows the steering torque with respect tosteering wheel angle when different steering feel gain is usedat various vehicle speeds It is observed that when the steeringfeel gain is increased at maximum range which is 1 thesteering torque will increase and driver will sense stiffer onthe steering wheel at high speed manoeuvre Meanwhilewhen decreasing the steering feel gain this will result inreducing the steering torqueThus the driver will sense softer
Target monitor
Host PC
Cross over cableConnector block
PCI 6221
Power supplyCurrent sensor
DC motor driver
Figure 14 Steering wheel system hardware overview
during turning the steering wheel at the low speed manoeu-vre However when the steering feel gain is 5 which exceedsits maximum limit the steering torque becomes unstableFigure 13(b) shows the steering feel gains at various speedsAt high speed of 120 kmh the steering feel gain is increasedand this will increase the steering torque for steering stabilityMeanwhile at low speed the steering feel gain is decreasedwhich will help improve vehicle manoeuvre especially duringparking Based on the proposed gain scheduling with theLQR controller not only does it provide better control on thesteering torque but also it is optional for a designer to have awide range adjustment to vary the steering torque for a betterdriver steering feel and to improve vehicle manoeuvre
4 Hardware in the Loop (HIL) and Result
For hardware in the loop (HIL) setup the front axle systemand linear vehicle model are simulated on the host computerThe steering wheel system is replaced by HIL mechanismThe position of the steering wheel motor is measured usinga rotary encoder with 500 pulses per revolution To measurethe steering wheel torque the current sensor is connectedin series with the DC motor driver The steering wheel DCmotor has a torque 129Nm with 100 rpm The single tracklinear vehicle model is used to generate the estimation selfaligning torque as element to the feedback torque and distur-bance to the front axle system The HIL system is executedunder 1ms sampling time and interfaced using Matlab-XPCtarget toolbox configuration as shown in Figures 14 and 15
The assumptions through the experiment are listed asfollows
(i) the vehicle cruises at constant speed(ii) the condition of dry asphalt road (120583 = 085)
The experiments are done under high medium and lowspeed manoeuvre [26] The same input steering wheel anglebetween simulation and experiment is given Therefore theresult obtained from real time control is compared with theelectronic power steering (EPS) system [15]
41 Motor Model Validation for Steering Wheel SystemAlthough the model of steering wheel system has been
12 International Journal of Vehicular Technology
Stee
ring
whe
el sy
stem
DC motor
Current
Encoder
16 bit card PCI 6221
Analogoutput
Analoginput
Counterchannel
TargetPC
Monitortarget
PC
Crossover cable
Host PC
Front axle systemand
linear vehicle model
Hardware part Simulation part
Figure 15 Block diagram of the steering wheel-XPC target interfaced
Steering wheel model
120575m1
120575m1
ia1
r(t)
Figure 16 Block diagrammotormodel validation for steeringwheelsystem
derived the model validation is necessary to ensure thesystem is fairly able to adequate for real time applicationThis section describes the motor model validation of steeringwheel system The block diagram is shown in Figure 16
The input signal is 119903(119905) and the output is rate change ofmotor ( 120575
1198981) motor position (1205751198981) and the motor current
( 1198941198981) The measured output response is then used and com-
pared with simulation to estimate the parameter of the steer-ing wheel model using the Matlab tool software The mea-sured output data are taken based on the open loop systemwhen the input to the system is sinusoidal signal withamplitude of plusmn45∘ for 10 sec The Gaussian method is usedto estimate the parameters of the steering wheel motorThe comparison between simulation and measured outputis shown in Figure 17 It is obtained that the output of thesimulation is almost identical to the measured output How-ever small difference errors occur due to the small parametricuncertainties that have not been modelled The parameterestimation value is illustrated in Table 6
42 Noise Filtering Both analog and digital devices have atrait of noise or unwanted features that have effect on thesystem performance such as electromagnetic interferenceThe noise could be either random or white noise Attenuationof this noise is often as a primary goal in control systemapplication and in particular when controlling the processsystem The average value of noise typically is zero whichwill give misreading information to the process and it wouldnot be possible to control the process In order to reduce thenoise inference filtering technique is used A filter is used toremove unwanted noise It removes some noise frequenciesto suppress interfacing signal and reduce background noise
Table 6 Parameter estimation for DCmotor steering wheel system
Parameter Value Estimation value Unit1198771 4405 564 ohm1198711 0010 0017 henry119870119904119898
1459 0024 Nm1198691198981 0002 00036 (kgm2)1198871198981 0004 00068 Nm(rads)1198961199041 165 0025 Nm
There are many types of filtering techniques such as low passfilter that has been used in this study The low pass filter isused to filter out the interference noise from the output of thecurrent sensor
Figure 18 shows the comparison of the output responsefor a current sensor with and without using filters Withouta filter the current sensor has high amplitude interferencenoise However with low pass filter under the frequency of5Hz the interference noise is reduced and the fluctuations areremoved Thus this provides a smoother output response
43 Effect on Compensation Torque A steering torque com-parison without and with compensation torque is shown inFigure 19The results for without compensation torque showsthat the steering torque response curve is far away fromEPS steering system with different amplitude as shown inFigure 19(a) This happens due to less of derivative reactiontorque from the steering wheel motor Figure 19(b) showssteering torque response with compensation torque
It is observed that by adding the additional compensationtorque the steering torque response curve trends fairly closeto EPS steering system Thus it helps to improve the driversteering feel during manoeuvre and stabilize the systemHowever with the wide range adjustment gains of the com-pensation torque it could improve the steering torqueresponse
44 Medium Speed Manoeuvre The experimental steeringtorque results at medium speed manoeuvre are shown inFigure 20(b) when the steering wheel angle is plusmn100∘ as shown
International Journal of Vehicular Technology 13
0 2 4 6 8 10minus50
0
50
Time (s)
Rate
chan
ge m
otor
pos
ition
(deg
s)
MeasurementSimulation
(a) Rate change of motor position
0 2 4 6 8 100
10
20
30
40
50
60
Time (s)
Mot
or p
ositi
on (d
eg)
Measurement
Simulation
(b) Position of the motor
0 2 4 6 8 10minus01
minus005
0
005
01
015
Mot
or cu
rren
t (A
)
MeasurementSimulation
Time (s)
(c) Current of the motor
Figure 17 Motor model validation comparison between simulation and measurement
With filter Without filter
Time (s)
Curr
ent (
A)
008
006
004
002
0
minus002
minus004
minus006
minus0081 15 2 25 3 35 4 45 5 55
Figure 18 Noise filtering using low pass filter
in Figure 20(a) From the results the experimental steeringtorque response is almost identical to the EPS steering systemHowever it has small different torque amplitudes These
occur due to the fact that mechanism or friction from themechanical gearbox motor is not modelled in proposedcontrol algorithm
14 International Journal of Vehicular Technology
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
minus152 3 4 5 6 7 8 9 10
(a) Without compensation torque
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
21 3 4 5 6 7 8 9 10
(b) With compensation torque
Figure 19 Steering torque influence on compensation torque
minus80
minus60
minus40
minus20
0
20
40
60
80
100
Stee
ring
whe
el an
gle (
deg)
Time (s)3 4 5 6 7 8 9 10
(a) Steering wheel angle at 80 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
15
20
5
0
minus5
minus10
minus15
minus203 4 5 6 7 8 9 10
(b) Steering torque estimation at 80 kmh
Figure 20 Steering torque estimation and steering wheel angle response at 80 kmh
45 High Speed Manoeuvre At high speed manoeuvre theinput steering wheel angle that is more than plusmn45∘ is very dan-gerous [12] Therefore in this experiment the input steeringwheel angle is less than plusmn45∘ as shown in Figure 21(a) Basedon the results shown in Figures 21(b) and 21(c) the experi-mental steering torque is obtained fairly similar to the EPSsteering system Although the steering wheel angle is at smallangles the influence of vehicle speed cause the steering torqueis increased Moreover at higher speed the vehicle has moreunder steer propensity [9] This increased the self aligningtorque in proportional high torque at front axle motorThusthe driver will sense stiffer on the steering wheel duringmanoeuvre
In general the centering of the steering wheel perfor-mance or steering wheel returnability is better when thesteering torque gradient is large and the hysteresis is small [9]Based on this fact the result is shown in Figure 21(c) whereby
the hysteresis of the steering wheel returnability based onexperimental result is smaller compared to EPS steeringsystem In addition the steering torque gradient is larger thanEPS steering torque Therefore it is considered that the pro-posed feedback torque algorithmnot only achieved a steeringfeel but also improved the steering wheel returnability
46 Lower Speed Manoeuvre The steering torque results atlow speed manoeuvre are shown in Figures 22(b) and 22(c)when the input steering wheel is as shown in Figure 22(a)The vehicle cruised at constant speed of 10 kmh
The experimental steering torque is almost comparableto EPS torque but the amplitude is decreases The torque isdecreased due to the fact that the vehicle speed is reducedwhich is effect on reducing the demand to the total feedbacktorque Moreover the large turning of the steering wheelincreases the backlash of the motor which then contributes
International Journal of Vehicular Technology 15
2 3 4 5 6 7 8 9 10minus30
minus20
minus10
0
10
20
30
40
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle at 120 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steeringSimulationExperiment
10
5
0
minus5
minus10
3 4 5 6 7 8 9
(b) Steering torque estimation at 120 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment10
15
5
0
minus5
minus10
minus15minus20 minus10 0 10 20 30 40
Steering wheel angle (deg)
(c) Steering torque estimation against steering wheel angle at 120 kmh
Figure 21 Steering torque estimation response at 120 kmh
to the decrease of the motor torque Thus driver will sensesofter during steering the wheel during maneuver Howeverit has small different torque against EPS steering Wide rangeadjustment of gains in proposed control algorithm may helpimprove the steering torque response
3 4 5 6 7 8 9 10minus150
minus100
minus50
0
50
100
150
200
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
10
15
5
0
minus5
minus10
minus15
Time (s)2 3 4 5 6 7 8 9 10
(b) Steering torque estimation at 10 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
Steering wheel angle (deg)
10
8
6
4
2
0
minus2
minus4
minus6
minus8
minus10
minus100 minus50 0 50 100 150
(c) Steering torque estimation against steering wheel angle at 10 kmh
Figure 22 Steering torque estimation response at 10 kmh
5 Conclusion and Future Work
The proposed force feedback torque estimation and controlalgorithm have been described The proposed control algo-rithm is considered as an effort to generate the feedback
16 International Journal of Vehicular Technology
torque for driver steering feel in the vehicle SBW systemThe mathematical model of SBW subsystem consists ofsteering wheel and front axle systems that are modelledThe effectiveness of the control algorithm is verified throughthe experimental hardware in the loop (HIL) interfacedwith Matlab XPC target toolbox software According to theexperimental results the proposed control algorithm is ableto generate the feedback torque and the driver is able to sensesteering feel similar to that obtained in EPS steering systemThe LQR controller with gain scheduling based on steeringwheel angle and vehicle speed function provides better torquecontrol that allows rejection of the uncertainty torque froma road condition Furthermore the steering feel gain is ableto increase and lower the steering torque with improvedsteering stability at high speed and it improved the vehiclemanoeuvre at low speed It is also found that compensationtorque is able to improve the feedback torque and to stabilizethe system However it has small different torque magnitudethat occurs due to the backlash of steering wheel motor anduncertainty disturbance of mechanical properties Howeverwith a wide range adjustment of the gains on the proposedcontrol algorithm it may improve and provide realisticsteering torque for better driver steering feel
In future work the proposed control algorithm will betested in real vehicle in order to have more realistic steeringtorque for further investigation and analysis
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
Acknowledgments
This work is fully supported by the Ministry of Higher Edu-cation Malaysia and University Technology Malaysia underFRGS (vote 4F370) and Research University Grant (vote00G64) This work is also supported by Proton HoldingBerhad
References
[1] Y Yao ldquoVehicle steer-by-wire system controlrdquo SAE TechnicalPaper 2006-01-1175 2006
[2] E A Kumar D Dinesh and N Kamble ldquoAn overview of activefront steering systemrdquo International Journal of Scientific ampEngineering Research vol 3 no 6 2012
[3] S M H Fahami H Zamzuri S A Mazlan and N B Zulkar-nain ldquoThe design of vehicle active front steering based on steerby wire systemrdquo Advanced Science Letters vol 19 no 1 pp 61ndash65 2013
[4] S M H Fahami H Zamzuri S A Mazlan and S A SaruchildquoThe variable steering ratio for vehicle steer by wire systemusing hyperbolic tangent methodrdquo Applied Mechanics andMaterials vol 575 pp 781ndash784 2014
[5] T Kaufmann S Milsap B Murray and J Petrowski ldquoDevelop-ment experience with steer by wirerdquo in Proceedings of the SAEInternational Congress and Exhibition SAE 2001-01-2479 2001
[6] S Amberkare F Bolourchi J Demerly and S Millsap ldquoAcontrol system methodology for steer by wire systemsrdquo SAETechnical Paper 2004-01-1106 2004
[7] A Baviskar J R Wagner D M Dawson D Braganza andP Setlur ldquoAn adjustable steer-by-wire haptic-interface trackingcontroller for ground vehiclesrdquo IEEE Transactions on VehicularTechnology vol 58 no 2 pp 546ndash554 2009
[8] M Segawa S Kimura T Kada and S Nakona ldquoA study of reac-tive torque control for steer bywire systemrdquo inProceedings of theInternational Symposium on Advanced Vehicle Control (AVECrsquo02) pp 653ndash658 Hiroshima Japan September 2002
[9] S W Oh H C Chae S C Yun and C S Han ldquoThe design ofa controller for the steer-by-wire systemrdquo JSME InternationalJournal Series C Mechanical Systems Machine Elements andManufacturing vol 47 no 3 pp 896ndash907 2004
[10] C J Kim J H Jang S K Oh J Y Lee and J K HedrickldquoDevelopment of a control algorithm for a rack-actuating steer-by-wire system using road information feedbackrdquo Proceedingsof the Institution of Mechanical Engineers Part D Journal ofAutomobile Engineering vol 222 no 9 pp 1559ndash1571 2008
[11] S Asai H Kuroyanagi and S Takeuchi ldquoDevelopment of asteer-by-wire system with force feedback using a disturbanceobserverrdquo inProceedings of the SAEWorldCongressampExhibitionon Steering amp Suspension Technology Detroit Mich USA 2004
[12] T J Park C S Han and S H Lee ldquoDevelopment of the elec-tronic control unit for the rack-actuating steer-by-wire usingthe hardware-in-the-loop simulation systemrdquo Mechatronicsvol 15 no 8 pp 899ndash918 2005
[13] D Odenhtal T Bunte H D Heitzer and C Eicker ldquoHow tomake steer-by-wire feel like power steeringrdquo in Proceedings ofthe 15th IFACWorld Congress Barcelona Spain July 2000
[14] H Na C Zong and D Hu ldquoInvestigations on corneringcontrol algorithm design and road feeling optimization for asteer-by-wire vehiclerdquo in Proceedings of the IEEE InternationalConference on Mechatronics and Automation (ICMA rsquo09) pp3246ndash3251 Changchun China August 2009
[15] S Ancha A Baviskar J RWagner andDMDawson ldquoGroundvehicle steering systems modelling control and analysis ofhydraulic electric and steer-by-wire configurationsrdquo Interna-tional Journal of Vehicle Design vol 44 no 1-2 pp 188ndash2082007
[16] P Yih and J C Gerdes ldquoModification of vehicle handlingcharacteristics via steer-by-wirerdquo IEEE Transactions on ControlSystems Technology vol 13 no 6 pp 965ndash976 2005
[17] S M H Fahami H Zamzuri S A Mazlan and M A ZakarialdquoModeling and simulation of vehicle steer by wire systemrdquoin Proceedings of the IEEE Symposium on Humanities Scienceand Engineering Research (SHUSER rsquo12) pp 765ndash770 KualaLumpur Malaysia June 2012
[18] S Anwar and L Chen ldquoAn analytical redundancy-based faultdetection and isolation algorithm for a road-wheel controlsubsystem in a steer-by-wire systemrdquo IEEE Transactions onVehicular Technology vol 56 no 5 pp 2859ndash2869 2007
[19] M T Do Z Man C Zhang H Wang and F S Tay ldquoRobustsliding mode-based learning control for steer-by-wire systemsin modern vehiclesrdquo IEEE Transactions on Vehicular Technol-ogy vol 63 no 2 pp 580ndash590 2014
[20] J Duan R Wang and Y Yu ldquoResearch on control strategies ofsteer-by-wire systemrdquo in Proceedings of the International Con-ference on Intelligent Computation Technology and Automation(ICICTA rsquo10) vol 2 pp 1122ndash1125 May 2010
International Journal of Vehicular Technology 17
[21] B Zheng and S Anwar ldquoYaw stability control of a steer-by-wireequipped vehicle via active front wheel steeringrdquoMechatronicsvol 19 no 6 pp 799ndash804 2009
[22] S Sahoo S C Subramanian and S Srivastava ldquoDesign andimplementation of a controller for navigating an autonomousground vehiclerdquo in Proceedings of the 2nd International Confer-ence on Power Control and Embedded Systems (ICPCES rsquo12) pp1ndash6 IEEE Allahabad India December 2012
[23] T D Gillespie Fundamentals of Vehicle Dynamics Society ofAutomotive Engineers 1992
[24] M H Lee S K Ha J Y Choi and K S Yoon ldquoImprovement ofthe steering feel of an electric power steering system by torquemap modificationrdquo Journal of Mechanical Science and Technol-ogy vol 19 no 3 pp 792ndash801 2005
[25] R Pastorino M A Naya J A Perez and J Cuadrado ldquoGearedPM coreless motor modelling for driverrsquos force feedback insteer-by-wire systemsrdquo Mechatronics vol 21 no 6 pp 1043ndash1054 2011
[26] P Koehn andM Eckrich ldquoActive steeringmdashthe BMWapproachtowards modern steering technologyrdquo SAE Technical PaperSeries 2004-01-1105 SAE International 2004
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
10 International Journal of Vehicular Technology
0 2 4 6 8 10minus50
minus40
minus30
minus20
minus10
0
10
20
30
40
50
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Input steering wheel angle
0 1 2 3 4 5 6 7 8 9 10minus15
minus10
minus5
0
5
10
Time (s)
Stee
ring
torq
ue (N
m) LQR
LQR + GS
Reference
(b) 120583 = 1
0 1 2 3 4 5 6 7 8 9 10
minus15
minus10
minus5
0
5
10
15
Stee
ring
torq
ue (N
m) LQR
Reference
Time (s)
LQR + GS
(c) 120583 = 14
0 1 2 3 4 5 6 7 8 9 10minus20
minus15
minus10
minus5
0
5
10
15
20
Stee
ring
torq
ue (N
m) LQR LQR + GS
Reference
Time (s)
(d) 120583 = 17
Figure 12 Comparison of torque response of LQR and LQR + GS controller at different 120583
Table 4 The gains of LQR controller and steering feel
119881 (kmh) 120575119904119908
(degree) 1198961 1198962 1198963 119896feel
(0ndash30) (plusmn180∘ lt 120575119904119908
lt plusmn405∘) 003 065 0001 014(31ndash100) (plusmn46∘ lt 120575
119904119908lt plusmn179∘) 003 084 0001 060
(101ndash120) 120575119904119908
le +45∘ or 120575119904119908
le minus45∘ 005 095 0001 095
angle is also divided into 3 categories according to vehiclespeed condition The notation ldquo+rdquo means that the driverturned the steering wheel clockwise and notation ldquominusrdquo is usedfor counter clockwise in which the maximum steering wheelturn is plusmn405∘
The gains of the LQR controller are defined using Brysonrsquosrule method and the steering feel gain is defined usingintuitive and has a range from 01 to 1 This range has beendefined for various speeds 10 kmh 80 kmh and 120 kmhafter comparing with EPS steering system [15] For exampleif the driver turns a steering wheel between (plusmn180∘ lt 120575
119904119908lt
plusmn405∘) and vehicle speed is 10 kmh the total feedback torque
is multiplied by steering feel gain 119896feel = 014 Thus thefeedback torque is decreased which results in making thedriver feel it easier turning the steering wheel and improvesvehicle manoeuvre The feedback torque is regulated by LQRcontroller when the gains are 1198961 = 003 1198962 = 065 and1198963 = 0001 The gains are the optimum that has been assuredby providing a better torque control based on vehicle speedand steering wheel angle function
Comparing torque response between LQR and LQR +GScontroller for different coefficient (120583) is shown in Figure 12when the input steering wheel angle is a double lane changemanoeuvre as shown in Figure 12(a) The vehicle speed isset to be 80 kmh Based on results obtained the LQR + GStechnique offers an improvement 90 and the curve trend isalmost similar to reference torque compared to the LQR con-troller This is because the gains of LQR controller changedbased on the steering wheel angle and vehicle speed param-eter which is able to reject the uncertainty torque of a roadcondition Meanwhile with the fixed gains of the LQR con-troller it has an excessive torque that can cause the system tobe unstable The RMS torque values are illustrated in Table 5
International Journal of Vehicular Technology 11
Table 5 Comparison of RMS torque values of LQR and LQR + GScontroller
Reference LQR LQR + GS120583 = 1 56361 50795 55450120583 = 14 61218 68339 57243120583 = 17 66184 72888 62080
minus50 0 50minus15
minus10
minus5
0
5
10
15
Steering angle (deg)
Stee
ring
torq
ue (N
m) Feel gain = 5
Feel gain = 01Feel gain = 05
Feel gain = 1
(a) Different steering feel gain at 80 kmh
minus50 minus40 minus30 minus20 minus10 0 10 20 30 40 50
minus15
minus10
minus5
0
5
10
15
Steering wheel angle (deg)
Stee
ring
torq
ue (N
m)
80kmh kfeel = 060
120 kmh kfeel = 095
10kmh kfeel = 014
(b) Different steering feel gain at various speed
Figure 13 Steering feel gains response at various vehicle speed
Figure 13(a) shows the steering torque with respect tosteering wheel angle when different steering feel gain is usedat various vehicle speeds It is observed that when the steeringfeel gain is increased at maximum range which is 1 thesteering torque will increase and driver will sense stiffer onthe steering wheel at high speed manoeuvre Meanwhilewhen decreasing the steering feel gain this will result inreducing the steering torqueThus the driver will sense softer
Target monitor
Host PC
Cross over cableConnector block
PCI 6221
Power supplyCurrent sensor
DC motor driver
Figure 14 Steering wheel system hardware overview
during turning the steering wheel at the low speed manoeu-vre However when the steering feel gain is 5 which exceedsits maximum limit the steering torque becomes unstableFigure 13(b) shows the steering feel gains at various speedsAt high speed of 120 kmh the steering feel gain is increasedand this will increase the steering torque for steering stabilityMeanwhile at low speed the steering feel gain is decreasedwhich will help improve vehicle manoeuvre especially duringparking Based on the proposed gain scheduling with theLQR controller not only does it provide better control on thesteering torque but also it is optional for a designer to have awide range adjustment to vary the steering torque for a betterdriver steering feel and to improve vehicle manoeuvre
4 Hardware in the Loop (HIL) and Result
For hardware in the loop (HIL) setup the front axle systemand linear vehicle model are simulated on the host computerThe steering wheel system is replaced by HIL mechanismThe position of the steering wheel motor is measured usinga rotary encoder with 500 pulses per revolution To measurethe steering wheel torque the current sensor is connectedin series with the DC motor driver The steering wheel DCmotor has a torque 129Nm with 100 rpm The single tracklinear vehicle model is used to generate the estimation selfaligning torque as element to the feedback torque and distur-bance to the front axle system The HIL system is executedunder 1ms sampling time and interfaced using Matlab-XPCtarget toolbox configuration as shown in Figures 14 and 15
The assumptions through the experiment are listed asfollows
(i) the vehicle cruises at constant speed(ii) the condition of dry asphalt road (120583 = 085)
The experiments are done under high medium and lowspeed manoeuvre [26] The same input steering wheel anglebetween simulation and experiment is given Therefore theresult obtained from real time control is compared with theelectronic power steering (EPS) system [15]
41 Motor Model Validation for Steering Wheel SystemAlthough the model of steering wheel system has been
12 International Journal of Vehicular Technology
Stee
ring
whe
el sy
stem
DC motor
Current
Encoder
16 bit card PCI 6221
Analogoutput
Analoginput
Counterchannel
TargetPC
Monitortarget
PC
Crossover cable
Host PC
Front axle systemand
linear vehicle model
Hardware part Simulation part
Figure 15 Block diagram of the steering wheel-XPC target interfaced
Steering wheel model
120575m1
120575m1
ia1
r(t)
Figure 16 Block diagrammotormodel validation for steeringwheelsystem
derived the model validation is necessary to ensure thesystem is fairly able to adequate for real time applicationThis section describes the motor model validation of steeringwheel system The block diagram is shown in Figure 16
The input signal is 119903(119905) and the output is rate change ofmotor ( 120575
1198981) motor position (1205751198981) and the motor current
( 1198941198981) The measured output response is then used and com-
pared with simulation to estimate the parameter of the steer-ing wheel model using the Matlab tool software The mea-sured output data are taken based on the open loop systemwhen the input to the system is sinusoidal signal withamplitude of plusmn45∘ for 10 sec The Gaussian method is usedto estimate the parameters of the steering wheel motorThe comparison between simulation and measured outputis shown in Figure 17 It is obtained that the output of thesimulation is almost identical to the measured output How-ever small difference errors occur due to the small parametricuncertainties that have not been modelled The parameterestimation value is illustrated in Table 6
42 Noise Filtering Both analog and digital devices have atrait of noise or unwanted features that have effect on thesystem performance such as electromagnetic interferenceThe noise could be either random or white noise Attenuationof this noise is often as a primary goal in control systemapplication and in particular when controlling the processsystem The average value of noise typically is zero whichwill give misreading information to the process and it wouldnot be possible to control the process In order to reduce thenoise inference filtering technique is used A filter is used toremove unwanted noise It removes some noise frequenciesto suppress interfacing signal and reduce background noise
Table 6 Parameter estimation for DCmotor steering wheel system
Parameter Value Estimation value Unit1198771 4405 564 ohm1198711 0010 0017 henry119870119904119898
1459 0024 Nm1198691198981 0002 00036 (kgm2)1198871198981 0004 00068 Nm(rads)1198961199041 165 0025 Nm
There are many types of filtering techniques such as low passfilter that has been used in this study The low pass filter isused to filter out the interference noise from the output of thecurrent sensor
Figure 18 shows the comparison of the output responsefor a current sensor with and without using filters Withouta filter the current sensor has high amplitude interferencenoise However with low pass filter under the frequency of5Hz the interference noise is reduced and the fluctuations areremoved Thus this provides a smoother output response
43 Effect on Compensation Torque A steering torque com-parison without and with compensation torque is shown inFigure 19The results for without compensation torque showsthat the steering torque response curve is far away fromEPS steering system with different amplitude as shown inFigure 19(a) This happens due to less of derivative reactiontorque from the steering wheel motor Figure 19(b) showssteering torque response with compensation torque
It is observed that by adding the additional compensationtorque the steering torque response curve trends fairly closeto EPS steering system Thus it helps to improve the driversteering feel during manoeuvre and stabilize the systemHowever with the wide range adjustment gains of the com-pensation torque it could improve the steering torqueresponse
44 Medium Speed Manoeuvre The experimental steeringtorque results at medium speed manoeuvre are shown inFigure 20(b) when the steering wheel angle is plusmn100∘ as shown
International Journal of Vehicular Technology 13
0 2 4 6 8 10minus50
0
50
Time (s)
Rate
chan
ge m
otor
pos
ition
(deg
s)
MeasurementSimulation
(a) Rate change of motor position
0 2 4 6 8 100
10
20
30
40
50
60
Time (s)
Mot
or p
ositi
on (d
eg)
Measurement
Simulation
(b) Position of the motor
0 2 4 6 8 10minus01
minus005
0
005
01
015
Mot
or cu
rren
t (A
)
MeasurementSimulation
Time (s)
(c) Current of the motor
Figure 17 Motor model validation comparison between simulation and measurement
With filter Without filter
Time (s)
Curr
ent (
A)
008
006
004
002
0
minus002
minus004
minus006
minus0081 15 2 25 3 35 4 45 5 55
Figure 18 Noise filtering using low pass filter
in Figure 20(a) From the results the experimental steeringtorque response is almost identical to the EPS steering systemHowever it has small different torque amplitudes These
occur due to the fact that mechanism or friction from themechanical gearbox motor is not modelled in proposedcontrol algorithm
14 International Journal of Vehicular Technology
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
minus152 3 4 5 6 7 8 9 10
(a) Without compensation torque
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
21 3 4 5 6 7 8 9 10
(b) With compensation torque
Figure 19 Steering torque influence on compensation torque
minus80
minus60
minus40
minus20
0
20
40
60
80
100
Stee
ring
whe
el an
gle (
deg)
Time (s)3 4 5 6 7 8 9 10
(a) Steering wheel angle at 80 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
15
20
5
0
minus5
minus10
minus15
minus203 4 5 6 7 8 9 10
(b) Steering torque estimation at 80 kmh
Figure 20 Steering torque estimation and steering wheel angle response at 80 kmh
45 High Speed Manoeuvre At high speed manoeuvre theinput steering wheel angle that is more than plusmn45∘ is very dan-gerous [12] Therefore in this experiment the input steeringwheel angle is less than plusmn45∘ as shown in Figure 21(a) Basedon the results shown in Figures 21(b) and 21(c) the experi-mental steering torque is obtained fairly similar to the EPSsteering system Although the steering wheel angle is at smallangles the influence of vehicle speed cause the steering torqueis increased Moreover at higher speed the vehicle has moreunder steer propensity [9] This increased the self aligningtorque in proportional high torque at front axle motorThusthe driver will sense stiffer on the steering wheel duringmanoeuvre
In general the centering of the steering wheel perfor-mance or steering wheel returnability is better when thesteering torque gradient is large and the hysteresis is small [9]Based on this fact the result is shown in Figure 21(c) whereby
the hysteresis of the steering wheel returnability based onexperimental result is smaller compared to EPS steeringsystem In addition the steering torque gradient is larger thanEPS steering torque Therefore it is considered that the pro-posed feedback torque algorithmnot only achieved a steeringfeel but also improved the steering wheel returnability
46 Lower Speed Manoeuvre The steering torque results atlow speed manoeuvre are shown in Figures 22(b) and 22(c)when the input steering wheel is as shown in Figure 22(a)The vehicle cruised at constant speed of 10 kmh
The experimental steering torque is almost comparableto EPS torque but the amplitude is decreases The torque isdecreased due to the fact that the vehicle speed is reducedwhich is effect on reducing the demand to the total feedbacktorque Moreover the large turning of the steering wheelincreases the backlash of the motor which then contributes
International Journal of Vehicular Technology 15
2 3 4 5 6 7 8 9 10minus30
minus20
minus10
0
10
20
30
40
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle at 120 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steeringSimulationExperiment
10
5
0
minus5
minus10
3 4 5 6 7 8 9
(b) Steering torque estimation at 120 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment10
15
5
0
minus5
minus10
minus15minus20 minus10 0 10 20 30 40
Steering wheel angle (deg)
(c) Steering torque estimation against steering wheel angle at 120 kmh
Figure 21 Steering torque estimation response at 120 kmh
to the decrease of the motor torque Thus driver will sensesofter during steering the wheel during maneuver Howeverit has small different torque against EPS steering Wide rangeadjustment of gains in proposed control algorithm may helpimprove the steering torque response
3 4 5 6 7 8 9 10minus150
minus100
minus50
0
50
100
150
200
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
10
15
5
0
minus5
minus10
minus15
Time (s)2 3 4 5 6 7 8 9 10
(b) Steering torque estimation at 10 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
Steering wheel angle (deg)
10
8
6
4
2
0
minus2
minus4
minus6
minus8
minus10
minus100 minus50 0 50 100 150
(c) Steering torque estimation against steering wheel angle at 10 kmh
Figure 22 Steering torque estimation response at 10 kmh
5 Conclusion and Future Work
The proposed force feedback torque estimation and controlalgorithm have been described The proposed control algo-rithm is considered as an effort to generate the feedback
16 International Journal of Vehicular Technology
torque for driver steering feel in the vehicle SBW systemThe mathematical model of SBW subsystem consists ofsteering wheel and front axle systems that are modelledThe effectiveness of the control algorithm is verified throughthe experimental hardware in the loop (HIL) interfacedwith Matlab XPC target toolbox software According to theexperimental results the proposed control algorithm is ableto generate the feedback torque and the driver is able to sensesteering feel similar to that obtained in EPS steering systemThe LQR controller with gain scheduling based on steeringwheel angle and vehicle speed function provides better torquecontrol that allows rejection of the uncertainty torque froma road condition Furthermore the steering feel gain is ableto increase and lower the steering torque with improvedsteering stability at high speed and it improved the vehiclemanoeuvre at low speed It is also found that compensationtorque is able to improve the feedback torque and to stabilizethe system However it has small different torque magnitudethat occurs due to the backlash of steering wheel motor anduncertainty disturbance of mechanical properties Howeverwith a wide range adjustment of the gains on the proposedcontrol algorithm it may improve and provide realisticsteering torque for better driver steering feel
In future work the proposed control algorithm will betested in real vehicle in order to have more realistic steeringtorque for further investigation and analysis
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
Acknowledgments
This work is fully supported by the Ministry of Higher Edu-cation Malaysia and University Technology Malaysia underFRGS (vote 4F370) and Research University Grant (vote00G64) This work is also supported by Proton HoldingBerhad
References
[1] Y Yao ldquoVehicle steer-by-wire system controlrdquo SAE TechnicalPaper 2006-01-1175 2006
[2] E A Kumar D Dinesh and N Kamble ldquoAn overview of activefront steering systemrdquo International Journal of Scientific ampEngineering Research vol 3 no 6 2012
[3] S M H Fahami H Zamzuri S A Mazlan and N B Zulkar-nain ldquoThe design of vehicle active front steering based on steerby wire systemrdquo Advanced Science Letters vol 19 no 1 pp 61ndash65 2013
[4] S M H Fahami H Zamzuri S A Mazlan and S A SaruchildquoThe variable steering ratio for vehicle steer by wire systemusing hyperbolic tangent methodrdquo Applied Mechanics andMaterials vol 575 pp 781ndash784 2014
[5] T Kaufmann S Milsap B Murray and J Petrowski ldquoDevelop-ment experience with steer by wirerdquo in Proceedings of the SAEInternational Congress and Exhibition SAE 2001-01-2479 2001
[6] S Amberkare F Bolourchi J Demerly and S Millsap ldquoAcontrol system methodology for steer by wire systemsrdquo SAETechnical Paper 2004-01-1106 2004
[7] A Baviskar J R Wagner D M Dawson D Braganza andP Setlur ldquoAn adjustable steer-by-wire haptic-interface trackingcontroller for ground vehiclesrdquo IEEE Transactions on VehicularTechnology vol 58 no 2 pp 546ndash554 2009
[8] M Segawa S Kimura T Kada and S Nakona ldquoA study of reac-tive torque control for steer bywire systemrdquo inProceedings of theInternational Symposium on Advanced Vehicle Control (AVECrsquo02) pp 653ndash658 Hiroshima Japan September 2002
[9] S W Oh H C Chae S C Yun and C S Han ldquoThe design ofa controller for the steer-by-wire systemrdquo JSME InternationalJournal Series C Mechanical Systems Machine Elements andManufacturing vol 47 no 3 pp 896ndash907 2004
[10] C J Kim J H Jang S K Oh J Y Lee and J K HedrickldquoDevelopment of a control algorithm for a rack-actuating steer-by-wire system using road information feedbackrdquo Proceedingsof the Institution of Mechanical Engineers Part D Journal ofAutomobile Engineering vol 222 no 9 pp 1559ndash1571 2008
[11] S Asai H Kuroyanagi and S Takeuchi ldquoDevelopment of asteer-by-wire system with force feedback using a disturbanceobserverrdquo inProceedings of the SAEWorldCongressampExhibitionon Steering amp Suspension Technology Detroit Mich USA 2004
[12] T J Park C S Han and S H Lee ldquoDevelopment of the elec-tronic control unit for the rack-actuating steer-by-wire usingthe hardware-in-the-loop simulation systemrdquo Mechatronicsvol 15 no 8 pp 899ndash918 2005
[13] D Odenhtal T Bunte H D Heitzer and C Eicker ldquoHow tomake steer-by-wire feel like power steeringrdquo in Proceedings ofthe 15th IFACWorld Congress Barcelona Spain July 2000
[14] H Na C Zong and D Hu ldquoInvestigations on corneringcontrol algorithm design and road feeling optimization for asteer-by-wire vehiclerdquo in Proceedings of the IEEE InternationalConference on Mechatronics and Automation (ICMA rsquo09) pp3246ndash3251 Changchun China August 2009
[15] S Ancha A Baviskar J RWagner andDMDawson ldquoGroundvehicle steering systems modelling control and analysis ofhydraulic electric and steer-by-wire configurationsrdquo Interna-tional Journal of Vehicle Design vol 44 no 1-2 pp 188ndash2082007
[16] P Yih and J C Gerdes ldquoModification of vehicle handlingcharacteristics via steer-by-wirerdquo IEEE Transactions on ControlSystems Technology vol 13 no 6 pp 965ndash976 2005
[17] S M H Fahami H Zamzuri S A Mazlan and M A ZakarialdquoModeling and simulation of vehicle steer by wire systemrdquoin Proceedings of the IEEE Symposium on Humanities Scienceand Engineering Research (SHUSER rsquo12) pp 765ndash770 KualaLumpur Malaysia June 2012
[18] S Anwar and L Chen ldquoAn analytical redundancy-based faultdetection and isolation algorithm for a road-wheel controlsubsystem in a steer-by-wire systemrdquo IEEE Transactions onVehicular Technology vol 56 no 5 pp 2859ndash2869 2007
[19] M T Do Z Man C Zhang H Wang and F S Tay ldquoRobustsliding mode-based learning control for steer-by-wire systemsin modern vehiclesrdquo IEEE Transactions on Vehicular Technol-ogy vol 63 no 2 pp 580ndash590 2014
[20] J Duan R Wang and Y Yu ldquoResearch on control strategies ofsteer-by-wire systemrdquo in Proceedings of the International Con-ference on Intelligent Computation Technology and Automation(ICICTA rsquo10) vol 2 pp 1122ndash1125 May 2010
International Journal of Vehicular Technology 17
[21] B Zheng and S Anwar ldquoYaw stability control of a steer-by-wireequipped vehicle via active front wheel steeringrdquoMechatronicsvol 19 no 6 pp 799ndash804 2009
[22] S Sahoo S C Subramanian and S Srivastava ldquoDesign andimplementation of a controller for navigating an autonomousground vehiclerdquo in Proceedings of the 2nd International Confer-ence on Power Control and Embedded Systems (ICPCES rsquo12) pp1ndash6 IEEE Allahabad India December 2012
[23] T D Gillespie Fundamentals of Vehicle Dynamics Society ofAutomotive Engineers 1992
[24] M H Lee S K Ha J Y Choi and K S Yoon ldquoImprovement ofthe steering feel of an electric power steering system by torquemap modificationrdquo Journal of Mechanical Science and Technol-ogy vol 19 no 3 pp 792ndash801 2005
[25] R Pastorino M A Naya J A Perez and J Cuadrado ldquoGearedPM coreless motor modelling for driverrsquos force feedback insteer-by-wire systemsrdquo Mechatronics vol 21 no 6 pp 1043ndash1054 2011
[26] P Koehn andM Eckrich ldquoActive steeringmdashthe BMWapproachtowards modern steering technologyrdquo SAE Technical PaperSeries 2004-01-1105 SAE International 2004
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
International Journal of Vehicular Technology 11
Table 5 Comparison of RMS torque values of LQR and LQR + GScontroller
Reference LQR LQR + GS120583 = 1 56361 50795 55450120583 = 14 61218 68339 57243120583 = 17 66184 72888 62080
minus50 0 50minus15
minus10
minus5
0
5
10
15
Steering angle (deg)
Stee
ring
torq
ue (N
m) Feel gain = 5
Feel gain = 01Feel gain = 05
Feel gain = 1
(a) Different steering feel gain at 80 kmh
minus50 minus40 minus30 minus20 minus10 0 10 20 30 40 50
minus15
minus10
minus5
0
5
10
15
Steering wheel angle (deg)
Stee
ring
torq
ue (N
m)
80kmh kfeel = 060
120 kmh kfeel = 095
10kmh kfeel = 014
(b) Different steering feel gain at various speed
Figure 13 Steering feel gains response at various vehicle speed
Figure 13(a) shows the steering torque with respect tosteering wheel angle when different steering feel gain is usedat various vehicle speeds It is observed that when the steeringfeel gain is increased at maximum range which is 1 thesteering torque will increase and driver will sense stiffer onthe steering wheel at high speed manoeuvre Meanwhilewhen decreasing the steering feel gain this will result inreducing the steering torqueThus the driver will sense softer
Target monitor
Host PC
Cross over cableConnector block
PCI 6221
Power supplyCurrent sensor
DC motor driver
Figure 14 Steering wheel system hardware overview
during turning the steering wheel at the low speed manoeu-vre However when the steering feel gain is 5 which exceedsits maximum limit the steering torque becomes unstableFigure 13(b) shows the steering feel gains at various speedsAt high speed of 120 kmh the steering feel gain is increasedand this will increase the steering torque for steering stabilityMeanwhile at low speed the steering feel gain is decreasedwhich will help improve vehicle manoeuvre especially duringparking Based on the proposed gain scheduling with theLQR controller not only does it provide better control on thesteering torque but also it is optional for a designer to have awide range adjustment to vary the steering torque for a betterdriver steering feel and to improve vehicle manoeuvre
4 Hardware in the Loop (HIL) and Result
For hardware in the loop (HIL) setup the front axle systemand linear vehicle model are simulated on the host computerThe steering wheel system is replaced by HIL mechanismThe position of the steering wheel motor is measured usinga rotary encoder with 500 pulses per revolution To measurethe steering wheel torque the current sensor is connectedin series with the DC motor driver The steering wheel DCmotor has a torque 129Nm with 100 rpm The single tracklinear vehicle model is used to generate the estimation selfaligning torque as element to the feedback torque and distur-bance to the front axle system The HIL system is executedunder 1ms sampling time and interfaced using Matlab-XPCtarget toolbox configuration as shown in Figures 14 and 15
The assumptions through the experiment are listed asfollows
(i) the vehicle cruises at constant speed(ii) the condition of dry asphalt road (120583 = 085)
The experiments are done under high medium and lowspeed manoeuvre [26] The same input steering wheel anglebetween simulation and experiment is given Therefore theresult obtained from real time control is compared with theelectronic power steering (EPS) system [15]
41 Motor Model Validation for Steering Wheel SystemAlthough the model of steering wheel system has been
12 International Journal of Vehicular Technology
Stee
ring
whe
el sy
stem
DC motor
Current
Encoder
16 bit card PCI 6221
Analogoutput
Analoginput
Counterchannel
TargetPC
Monitortarget
PC
Crossover cable
Host PC
Front axle systemand
linear vehicle model
Hardware part Simulation part
Figure 15 Block diagram of the steering wheel-XPC target interfaced
Steering wheel model
120575m1
120575m1
ia1
r(t)
Figure 16 Block diagrammotormodel validation for steeringwheelsystem
derived the model validation is necessary to ensure thesystem is fairly able to adequate for real time applicationThis section describes the motor model validation of steeringwheel system The block diagram is shown in Figure 16
The input signal is 119903(119905) and the output is rate change ofmotor ( 120575
1198981) motor position (1205751198981) and the motor current
( 1198941198981) The measured output response is then used and com-
pared with simulation to estimate the parameter of the steer-ing wheel model using the Matlab tool software The mea-sured output data are taken based on the open loop systemwhen the input to the system is sinusoidal signal withamplitude of plusmn45∘ for 10 sec The Gaussian method is usedto estimate the parameters of the steering wheel motorThe comparison between simulation and measured outputis shown in Figure 17 It is obtained that the output of thesimulation is almost identical to the measured output How-ever small difference errors occur due to the small parametricuncertainties that have not been modelled The parameterestimation value is illustrated in Table 6
42 Noise Filtering Both analog and digital devices have atrait of noise or unwanted features that have effect on thesystem performance such as electromagnetic interferenceThe noise could be either random or white noise Attenuationof this noise is often as a primary goal in control systemapplication and in particular when controlling the processsystem The average value of noise typically is zero whichwill give misreading information to the process and it wouldnot be possible to control the process In order to reduce thenoise inference filtering technique is used A filter is used toremove unwanted noise It removes some noise frequenciesto suppress interfacing signal and reduce background noise
Table 6 Parameter estimation for DCmotor steering wheel system
Parameter Value Estimation value Unit1198771 4405 564 ohm1198711 0010 0017 henry119870119904119898
1459 0024 Nm1198691198981 0002 00036 (kgm2)1198871198981 0004 00068 Nm(rads)1198961199041 165 0025 Nm
There are many types of filtering techniques such as low passfilter that has been used in this study The low pass filter isused to filter out the interference noise from the output of thecurrent sensor
Figure 18 shows the comparison of the output responsefor a current sensor with and without using filters Withouta filter the current sensor has high amplitude interferencenoise However with low pass filter under the frequency of5Hz the interference noise is reduced and the fluctuations areremoved Thus this provides a smoother output response
43 Effect on Compensation Torque A steering torque com-parison without and with compensation torque is shown inFigure 19The results for without compensation torque showsthat the steering torque response curve is far away fromEPS steering system with different amplitude as shown inFigure 19(a) This happens due to less of derivative reactiontorque from the steering wheel motor Figure 19(b) showssteering torque response with compensation torque
It is observed that by adding the additional compensationtorque the steering torque response curve trends fairly closeto EPS steering system Thus it helps to improve the driversteering feel during manoeuvre and stabilize the systemHowever with the wide range adjustment gains of the com-pensation torque it could improve the steering torqueresponse
44 Medium Speed Manoeuvre The experimental steeringtorque results at medium speed manoeuvre are shown inFigure 20(b) when the steering wheel angle is plusmn100∘ as shown
International Journal of Vehicular Technology 13
0 2 4 6 8 10minus50
0
50
Time (s)
Rate
chan
ge m
otor
pos
ition
(deg
s)
MeasurementSimulation
(a) Rate change of motor position
0 2 4 6 8 100
10
20
30
40
50
60
Time (s)
Mot
or p
ositi
on (d
eg)
Measurement
Simulation
(b) Position of the motor
0 2 4 6 8 10minus01
minus005
0
005
01
015
Mot
or cu
rren
t (A
)
MeasurementSimulation
Time (s)
(c) Current of the motor
Figure 17 Motor model validation comparison between simulation and measurement
With filter Without filter
Time (s)
Curr
ent (
A)
008
006
004
002
0
minus002
minus004
minus006
minus0081 15 2 25 3 35 4 45 5 55
Figure 18 Noise filtering using low pass filter
in Figure 20(a) From the results the experimental steeringtorque response is almost identical to the EPS steering systemHowever it has small different torque amplitudes These
occur due to the fact that mechanism or friction from themechanical gearbox motor is not modelled in proposedcontrol algorithm
14 International Journal of Vehicular Technology
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
minus152 3 4 5 6 7 8 9 10
(a) Without compensation torque
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
21 3 4 5 6 7 8 9 10
(b) With compensation torque
Figure 19 Steering torque influence on compensation torque
minus80
minus60
minus40
minus20
0
20
40
60
80
100
Stee
ring
whe
el an
gle (
deg)
Time (s)3 4 5 6 7 8 9 10
(a) Steering wheel angle at 80 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
15
20
5
0
minus5
minus10
minus15
minus203 4 5 6 7 8 9 10
(b) Steering torque estimation at 80 kmh
Figure 20 Steering torque estimation and steering wheel angle response at 80 kmh
45 High Speed Manoeuvre At high speed manoeuvre theinput steering wheel angle that is more than plusmn45∘ is very dan-gerous [12] Therefore in this experiment the input steeringwheel angle is less than plusmn45∘ as shown in Figure 21(a) Basedon the results shown in Figures 21(b) and 21(c) the experi-mental steering torque is obtained fairly similar to the EPSsteering system Although the steering wheel angle is at smallangles the influence of vehicle speed cause the steering torqueis increased Moreover at higher speed the vehicle has moreunder steer propensity [9] This increased the self aligningtorque in proportional high torque at front axle motorThusthe driver will sense stiffer on the steering wheel duringmanoeuvre
In general the centering of the steering wheel perfor-mance or steering wheel returnability is better when thesteering torque gradient is large and the hysteresis is small [9]Based on this fact the result is shown in Figure 21(c) whereby
the hysteresis of the steering wheel returnability based onexperimental result is smaller compared to EPS steeringsystem In addition the steering torque gradient is larger thanEPS steering torque Therefore it is considered that the pro-posed feedback torque algorithmnot only achieved a steeringfeel but also improved the steering wheel returnability
46 Lower Speed Manoeuvre The steering torque results atlow speed manoeuvre are shown in Figures 22(b) and 22(c)when the input steering wheel is as shown in Figure 22(a)The vehicle cruised at constant speed of 10 kmh
The experimental steering torque is almost comparableto EPS torque but the amplitude is decreases The torque isdecreased due to the fact that the vehicle speed is reducedwhich is effect on reducing the demand to the total feedbacktorque Moreover the large turning of the steering wheelincreases the backlash of the motor which then contributes
International Journal of Vehicular Technology 15
2 3 4 5 6 7 8 9 10minus30
minus20
minus10
0
10
20
30
40
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle at 120 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steeringSimulationExperiment
10
5
0
minus5
minus10
3 4 5 6 7 8 9
(b) Steering torque estimation at 120 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment10
15
5
0
minus5
minus10
minus15minus20 minus10 0 10 20 30 40
Steering wheel angle (deg)
(c) Steering torque estimation against steering wheel angle at 120 kmh
Figure 21 Steering torque estimation response at 120 kmh
to the decrease of the motor torque Thus driver will sensesofter during steering the wheel during maneuver Howeverit has small different torque against EPS steering Wide rangeadjustment of gains in proposed control algorithm may helpimprove the steering torque response
3 4 5 6 7 8 9 10minus150
minus100
minus50
0
50
100
150
200
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
10
15
5
0
minus5
minus10
minus15
Time (s)2 3 4 5 6 7 8 9 10
(b) Steering torque estimation at 10 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
Steering wheel angle (deg)
10
8
6
4
2
0
minus2
minus4
minus6
minus8
minus10
minus100 minus50 0 50 100 150
(c) Steering torque estimation against steering wheel angle at 10 kmh
Figure 22 Steering torque estimation response at 10 kmh
5 Conclusion and Future Work
The proposed force feedback torque estimation and controlalgorithm have been described The proposed control algo-rithm is considered as an effort to generate the feedback
16 International Journal of Vehicular Technology
torque for driver steering feel in the vehicle SBW systemThe mathematical model of SBW subsystem consists ofsteering wheel and front axle systems that are modelledThe effectiveness of the control algorithm is verified throughthe experimental hardware in the loop (HIL) interfacedwith Matlab XPC target toolbox software According to theexperimental results the proposed control algorithm is ableto generate the feedback torque and the driver is able to sensesteering feel similar to that obtained in EPS steering systemThe LQR controller with gain scheduling based on steeringwheel angle and vehicle speed function provides better torquecontrol that allows rejection of the uncertainty torque froma road condition Furthermore the steering feel gain is ableto increase and lower the steering torque with improvedsteering stability at high speed and it improved the vehiclemanoeuvre at low speed It is also found that compensationtorque is able to improve the feedback torque and to stabilizethe system However it has small different torque magnitudethat occurs due to the backlash of steering wheel motor anduncertainty disturbance of mechanical properties Howeverwith a wide range adjustment of the gains on the proposedcontrol algorithm it may improve and provide realisticsteering torque for better driver steering feel
In future work the proposed control algorithm will betested in real vehicle in order to have more realistic steeringtorque for further investigation and analysis
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
Acknowledgments
This work is fully supported by the Ministry of Higher Edu-cation Malaysia and University Technology Malaysia underFRGS (vote 4F370) and Research University Grant (vote00G64) This work is also supported by Proton HoldingBerhad
References
[1] Y Yao ldquoVehicle steer-by-wire system controlrdquo SAE TechnicalPaper 2006-01-1175 2006
[2] E A Kumar D Dinesh and N Kamble ldquoAn overview of activefront steering systemrdquo International Journal of Scientific ampEngineering Research vol 3 no 6 2012
[3] S M H Fahami H Zamzuri S A Mazlan and N B Zulkar-nain ldquoThe design of vehicle active front steering based on steerby wire systemrdquo Advanced Science Letters vol 19 no 1 pp 61ndash65 2013
[4] S M H Fahami H Zamzuri S A Mazlan and S A SaruchildquoThe variable steering ratio for vehicle steer by wire systemusing hyperbolic tangent methodrdquo Applied Mechanics andMaterials vol 575 pp 781ndash784 2014
[5] T Kaufmann S Milsap B Murray and J Petrowski ldquoDevelop-ment experience with steer by wirerdquo in Proceedings of the SAEInternational Congress and Exhibition SAE 2001-01-2479 2001
[6] S Amberkare F Bolourchi J Demerly and S Millsap ldquoAcontrol system methodology for steer by wire systemsrdquo SAETechnical Paper 2004-01-1106 2004
[7] A Baviskar J R Wagner D M Dawson D Braganza andP Setlur ldquoAn adjustable steer-by-wire haptic-interface trackingcontroller for ground vehiclesrdquo IEEE Transactions on VehicularTechnology vol 58 no 2 pp 546ndash554 2009
[8] M Segawa S Kimura T Kada and S Nakona ldquoA study of reac-tive torque control for steer bywire systemrdquo inProceedings of theInternational Symposium on Advanced Vehicle Control (AVECrsquo02) pp 653ndash658 Hiroshima Japan September 2002
[9] S W Oh H C Chae S C Yun and C S Han ldquoThe design ofa controller for the steer-by-wire systemrdquo JSME InternationalJournal Series C Mechanical Systems Machine Elements andManufacturing vol 47 no 3 pp 896ndash907 2004
[10] C J Kim J H Jang S K Oh J Y Lee and J K HedrickldquoDevelopment of a control algorithm for a rack-actuating steer-by-wire system using road information feedbackrdquo Proceedingsof the Institution of Mechanical Engineers Part D Journal ofAutomobile Engineering vol 222 no 9 pp 1559ndash1571 2008
[11] S Asai H Kuroyanagi and S Takeuchi ldquoDevelopment of asteer-by-wire system with force feedback using a disturbanceobserverrdquo inProceedings of the SAEWorldCongressampExhibitionon Steering amp Suspension Technology Detroit Mich USA 2004
[12] T J Park C S Han and S H Lee ldquoDevelopment of the elec-tronic control unit for the rack-actuating steer-by-wire usingthe hardware-in-the-loop simulation systemrdquo Mechatronicsvol 15 no 8 pp 899ndash918 2005
[13] D Odenhtal T Bunte H D Heitzer and C Eicker ldquoHow tomake steer-by-wire feel like power steeringrdquo in Proceedings ofthe 15th IFACWorld Congress Barcelona Spain July 2000
[14] H Na C Zong and D Hu ldquoInvestigations on corneringcontrol algorithm design and road feeling optimization for asteer-by-wire vehiclerdquo in Proceedings of the IEEE InternationalConference on Mechatronics and Automation (ICMA rsquo09) pp3246ndash3251 Changchun China August 2009
[15] S Ancha A Baviskar J RWagner andDMDawson ldquoGroundvehicle steering systems modelling control and analysis ofhydraulic electric and steer-by-wire configurationsrdquo Interna-tional Journal of Vehicle Design vol 44 no 1-2 pp 188ndash2082007
[16] P Yih and J C Gerdes ldquoModification of vehicle handlingcharacteristics via steer-by-wirerdquo IEEE Transactions on ControlSystems Technology vol 13 no 6 pp 965ndash976 2005
[17] S M H Fahami H Zamzuri S A Mazlan and M A ZakarialdquoModeling and simulation of vehicle steer by wire systemrdquoin Proceedings of the IEEE Symposium on Humanities Scienceand Engineering Research (SHUSER rsquo12) pp 765ndash770 KualaLumpur Malaysia June 2012
[18] S Anwar and L Chen ldquoAn analytical redundancy-based faultdetection and isolation algorithm for a road-wheel controlsubsystem in a steer-by-wire systemrdquo IEEE Transactions onVehicular Technology vol 56 no 5 pp 2859ndash2869 2007
[19] M T Do Z Man C Zhang H Wang and F S Tay ldquoRobustsliding mode-based learning control for steer-by-wire systemsin modern vehiclesrdquo IEEE Transactions on Vehicular Technol-ogy vol 63 no 2 pp 580ndash590 2014
[20] J Duan R Wang and Y Yu ldquoResearch on control strategies ofsteer-by-wire systemrdquo in Proceedings of the International Con-ference on Intelligent Computation Technology and Automation(ICICTA rsquo10) vol 2 pp 1122ndash1125 May 2010
International Journal of Vehicular Technology 17
[21] B Zheng and S Anwar ldquoYaw stability control of a steer-by-wireequipped vehicle via active front wheel steeringrdquoMechatronicsvol 19 no 6 pp 799ndash804 2009
[22] S Sahoo S C Subramanian and S Srivastava ldquoDesign andimplementation of a controller for navigating an autonomousground vehiclerdquo in Proceedings of the 2nd International Confer-ence on Power Control and Embedded Systems (ICPCES rsquo12) pp1ndash6 IEEE Allahabad India December 2012
[23] T D Gillespie Fundamentals of Vehicle Dynamics Society ofAutomotive Engineers 1992
[24] M H Lee S K Ha J Y Choi and K S Yoon ldquoImprovement ofthe steering feel of an electric power steering system by torquemap modificationrdquo Journal of Mechanical Science and Technol-ogy vol 19 no 3 pp 792ndash801 2005
[25] R Pastorino M A Naya J A Perez and J Cuadrado ldquoGearedPM coreless motor modelling for driverrsquos force feedback insteer-by-wire systemsrdquo Mechatronics vol 21 no 6 pp 1043ndash1054 2011
[26] P Koehn andM Eckrich ldquoActive steeringmdashthe BMWapproachtowards modern steering technologyrdquo SAE Technical PaperSeries 2004-01-1105 SAE International 2004
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
12 International Journal of Vehicular Technology
Stee
ring
whe
el sy
stem
DC motor
Current
Encoder
16 bit card PCI 6221
Analogoutput
Analoginput
Counterchannel
TargetPC
Monitortarget
PC
Crossover cable
Host PC
Front axle systemand
linear vehicle model
Hardware part Simulation part
Figure 15 Block diagram of the steering wheel-XPC target interfaced
Steering wheel model
120575m1
120575m1
ia1
r(t)
Figure 16 Block diagrammotormodel validation for steeringwheelsystem
derived the model validation is necessary to ensure thesystem is fairly able to adequate for real time applicationThis section describes the motor model validation of steeringwheel system The block diagram is shown in Figure 16
The input signal is 119903(119905) and the output is rate change ofmotor ( 120575
1198981) motor position (1205751198981) and the motor current
( 1198941198981) The measured output response is then used and com-
pared with simulation to estimate the parameter of the steer-ing wheel model using the Matlab tool software The mea-sured output data are taken based on the open loop systemwhen the input to the system is sinusoidal signal withamplitude of plusmn45∘ for 10 sec The Gaussian method is usedto estimate the parameters of the steering wheel motorThe comparison between simulation and measured outputis shown in Figure 17 It is obtained that the output of thesimulation is almost identical to the measured output How-ever small difference errors occur due to the small parametricuncertainties that have not been modelled The parameterestimation value is illustrated in Table 6
42 Noise Filtering Both analog and digital devices have atrait of noise or unwanted features that have effect on thesystem performance such as electromagnetic interferenceThe noise could be either random or white noise Attenuationof this noise is often as a primary goal in control systemapplication and in particular when controlling the processsystem The average value of noise typically is zero whichwill give misreading information to the process and it wouldnot be possible to control the process In order to reduce thenoise inference filtering technique is used A filter is used toremove unwanted noise It removes some noise frequenciesto suppress interfacing signal and reduce background noise
Table 6 Parameter estimation for DCmotor steering wheel system
Parameter Value Estimation value Unit1198771 4405 564 ohm1198711 0010 0017 henry119870119904119898
1459 0024 Nm1198691198981 0002 00036 (kgm2)1198871198981 0004 00068 Nm(rads)1198961199041 165 0025 Nm
There are many types of filtering techniques such as low passfilter that has been used in this study The low pass filter isused to filter out the interference noise from the output of thecurrent sensor
Figure 18 shows the comparison of the output responsefor a current sensor with and without using filters Withouta filter the current sensor has high amplitude interferencenoise However with low pass filter under the frequency of5Hz the interference noise is reduced and the fluctuations areremoved Thus this provides a smoother output response
43 Effect on Compensation Torque A steering torque com-parison without and with compensation torque is shown inFigure 19The results for without compensation torque showsthat the steering torque response curve is far away fromEPS steering system with different amplitude as shown inFigure 19(a) This happens due to less of derivative reactiontorque from the steering wheel motor Figure 19(b) showssteering torque response with compensation torque
It is observed that by adding the additional compensationtorque the steering torque response curve trends fairly closeto EPS steering system Thus it helps to improve the driversteering feel during manoeuvre and stabilize the systemHowever with the wide range adjustment gains of the com-pensation torque it could improve the steering torqueresponse
44 Medium Speed Manoeuvre The experimental steeringtorque results at medium speed manoeuvre are shown inFigure 20(b) when the steering wheel angle is plusmn100∘ as shown
International Journal of Vehicular Technology 13
0 2 4 6 8 10minus50
0
50
Time (s)
Rate
chan
ge m
otor
pos
ition
(deg
s)
MeasurementSimulation
(a) Rate change of motor position
0 2 4 6 8 100
10
20
30
40
50
60
Time (s)
Mot
or p
ositi
on (d
eg)
Measurement
Simulation
(b) Position of the motor
0 2 4 6 8 10minus01
minus005
0
005
01
015
Mot
or cu
rren
t (A
)
MeasurementSimulation
Time (s)
(c) Current of the motor
Figure 17 Motor model validation comparison between simulation and measurement
With filter Without filter
Time (s)
Curr
ent (
A)
008
006
004
002
0
minus002
minus004
minus006
minus0081 15 2 25 3 35 4 45 5 55
Figure 18 Noise filtering using low pass filter
in Figure 20(a) From the results the experimental steeringtorque response is almost identical to the EPS steering systemHowever it has small different torque amplitudes These
occur due to the fact that mechanism or friction from themechanical gearbox motor is not modelled in proposedcontrol algorithm
14 International Journal of Vehicular Technology
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
minus152 3 4 5 6 7 8 9 10
(a) Without compensation torque
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
21 3 4 5 6 7 8 9 10
(b) With compensation torque
Figure 19 Steering torque influence on compensation torque
minus80
minus60
minus40
minus20
0
20
40
60
80
100
Stee
ring
whe
el an
gle (
deg)
Time (s)3 4 5 6 7 8 9 10
(a) Steering wheel angle at 80 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
15
20
5
0
minus5
minus10
minus15
minus203 4 5 6 7 8 9 10
(b) Steering torque estimation at 80 kmh
Figure 20 Steering torque estimation and steering wheel angle response at 80 kmh
45 High Speed Manoeuvre At high speed manoeuvre theinput steering wheel angle that is more than plusmn45∘ is very dan-gerous [12] Therefore in this experiment the input steeringwheel angle is less than plusmn45∘ as shown in Figure 21(a) Basedon the results shown in Figures 21(b) and 21(c) the experi-mental steering torque is obtained fairly similar to the EPSsteering system Although the steering wheel angle is at smallangles the influence of vehicle speed cause the steering torqueis increased Moreover at higher speed the vehicle has moreunder steer propensity [9] This increased the self aligningtorque in proportional high torque at front axle motorThusthe driver will sense stiffer on the steering wheel duringmanoeuvre
In general the centering of the steering wheel perfor-mance or steering wheel returnability is better when thesteering torque gradient is large and the hysteresis is small [9]Based on this fact the result is shown in Figure 21(c) whereby
the hysteresis of the steering wheel returnability based onexperimental result is smaller compared to EPS steeringsystem In addition the steering torque gradient is larger thanEPS steering torque Therefore it is considered that the pro-posed feedback torque algorithmnot only achieved a steeringfeel but also improved the steering wheel returnability
46 Lower Speed Manoeuvre The steering torque results atlow speed manoeuvre are shown in Figures 22(b) and 22(c)when the input steering wheel is as shown in Figure 22(a)The vehicle cruised at constant speed of 10 kmh
The experimental steering torque is almost comparableto EPS torque but the amplitude is decreases The torque isdecreased due to the fact that the vehicle speed is reducedwhich is effect on reducing the demand to the total feedbacktorque Moreover the large turning of the steering wheelincreases the backlash of the motor which then contributes
International Journal of Vehicular Technology 15
2 3 4 5 6 7 8 9 10minus30
minus20
minus10
0
10
20
30
40
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle at 120 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steeringSimulationExperiment
10
5
0
minus5
minus10
3 4 5 6 7 8 9
(b) Steering torque estimation at 120 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment10
15
5
0
minus5
minus10
minus15minus20 minus10 0 10 20 30 40
Steering wheel angle (deg)
(c) Steering torque estimation against steering wheel angle at 120 kmh
Figure 21 Steering torque estimation response at 120 kmh
to the decrease of the motor torque Thus driver will sensesofter during steering the wheel during maneuver Howeverit has small different torque against EPS steering Wide rangeadjustment of gains in proposed control algorithm may helpimprove the steering torque response
3 4 5 6 7 8 9 10minus150
minus100
minus50
0
50
100
150
200
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
10
15
5
0
minus5
minus10
minus15
Time (s)2 3 4 5 6 7 8 9 10
(b) Steering torque estimation at 10 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
Steering wheel angle (deg)
10
8
6
4
2
0
minus2
minus4
minus6
minus8
minus10
minus100 minus50 0 50 100 150
(c) Steering torque estimation against steering wheel angle at 10 kmh
Figure 22 Steering torque estimation response at 10 kmh
5 Conclusion and Future Work
The proposed force feedback torque estimation and controlalgorithm have been described The proposed control algo-rithm is considered as an effort to generate the feedback
16 International Journal of Vehicular Technology
torque for driver steering feel in the vehicle SBW systemThe mathematical model of SBW subsystem consists ofsteering wheel and front axle systems that are modelledThe effectiveness of the control algorithm is verified throughthe experimental hardware in the loop (HIL) interfacedwith Matlab XPC target toolbox software According to theexperimental results the proposed control algorithm is ableto generate the feedback torque and the driver is able to sensesteering feel similar to that obtained in EPS steering systemThe LQR controller with gain scheduling based on steeringwheel angle and vehicle speed function provides better torquecontrol that allows rejection of the uncertainty torque froma road condition Furthermore the steering feel gain is ableto increase and lower the steering torque with improvedsteering stability at high speed and it improved the vehiclemanoeuvre at low speed It is also found that compensationtorque is able to improve the feedback torque and to stabilizethe system However it has small different torque magnitudethat occurs due to the backlash of steering wheel motor anduncertainty disturbance of mechanical properties Howeverwith a wide range adjustment of the gains on the proposedcontrol algorithm it may improve and provide realisticsteering torque for better driver steering feel
In future work the proposed control algorithm will betested in real vehicle in order to have more realistic steeringtorque for further investigation and analysis
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
Acknowledgments
This work is fully supported by the Ministry of Higher Edu-cation Malaysia and University Technology Malaysia underFRGS (vote 4F370) and Research University Grant (vote00G64) This work is also supported by Proton HoldingBerhad
References
[1] Y Yao ldquoVehicle steer-by-wire system controlrdquo SAE TechnicalPaper 2006-01-1175 2006
[2] E A Kumar D Dinesh and N Kamble ldquoAn overview of activefront steering systemrdquo International Journal of Scientific ampEngineering Research vol 3 no 6 2012
[3] S M H Fahami H Zamzuri S A Mazlan and N B Zulkar-nain ldquoThe design of vehicle active front steering based on steerby wire systemrdquo Advanced Science Letters vol 19 no 1 pp 61ndash65 2013
[4] S M H Fahami H Zamzuri S A Mazlan and S A SaruchildquoThe variable steering ratio for vehicle steer by wire systemusing hyperbolic tangent methodrdquo Applied Mechanics andMaterials vol 575 pp 781ndash784 2014
[5] T Kaufmann S Milsap B Murray and J Petrowski ldquoDevelop-ment experience with steer by wirerdquo in Proceedings of the SAEInternational Congress and Exhibition SAE 2001-01-2479 2001
[6] S Amberkare F Bolourchi J Demerly and S Millsap ldquoAcontrol system methodology for steer by wire systemsrdquo SAETechnical Paper 2004-01-1106 2004
[7] A Baviskar J R Wagner D M Dawson D Braganza andP Setlur ldquoAn adjustable steer-by-wire haptic-interface trackingcontroller for ground vehiclesrdquo IEEE Transactions on VehicularTechnology vol 58 no 2 pp 546ndash554 2009
[8] M Segawa S Kimura T Kada and S Nakona ldquoA study of reac-tive torque control for steer bywire systemrdquo inProceedings of theInternational Symposium on Advanced Vehicle Control (AVECrsquo02) pp 653ndash658 Hiroshima Japan September 2002
[9] S W Oh H C Chae S C Yun and C S Han ldquoThe design ofa controller for the steer-by-wire systemrdquo JSME InternationalJournal Series C Mechanical Systems Machine Elements andManufacturing vol 47 no 3 pp 896ndash907 2004
[10] C J Kim J H Jang S K Oh J Y Lee and J K HedrickldquoDevelopment of a control algorithm for a rack-actuating steer-by-wire system using road information feedbackrdquo Proceedingsof the Institution of Mechanical Engineers Part D Journal ofAutomobile Engineering vol 222 no 9 pp 1559ndash1571 2008
[11] S Asai H Kuroyanagi and S Takeuchi ldquoDevelopment of asteer-by-wire system with force feedback using a disturbanceobserverrdquo inProceedings of the SAEWorldCongressampExhibitionon Steering amp Suspension Technology Detroit Mich USA 2004
[12] T J Park C S Han and S H Lee ldquoDevelopment of the elec-tronic control unit for the rack-actuating steer-by-wire usingthe hardware-in-the-loop simulation systemrdquo Mechatronicsvol 15 no 8 pp 899ndash918 2005
[13] D Odenhtal T Bunte H D Heitzer and C Eicker ldquoHow tomake steer-by-wire feel like power steeringrdquo in Proceedings ofthe 15th IFACWorld Congress Barcelona Spain July 2000
[14] H Na C Zong and D Hu ldquoInvestigations on corneringcontrol algorithm design and road feeling optimization for asteer-by-wire vehiclerdquo in Proceedings of the IEEE InternationalConference on Mechatronics and Automation (ICMA rsquo09) pp3246ndash3251 Changchun China August 2009
[15] S Ancha A Baviskar J RWagner andDMDawson ldquoGroundvehicle steering systems modelling control and analysis ofhydraulic electric and steer-by-wire configurationsrdquo Interna-tional Journal of Vehicle Design vol 44 no 1-2 pp 188ndash2082007
[16] P Yih and J C Gerdes ldquoModification of vehicle handlingcharacteristics via steer-by-wirerdquo IEEE Transactions on ControlSystems Technology vol 13 no 6 pp 965ndash976 2005
[17] S M H Fahami H Zamzuri S A Mazlan and M A ZakarialdquoModeling and simulation of vehicle steer by wire systemrdquoin Proceedings of the IEEE Symposium on Humanities Scienceand Engineering Research (SHUSER rsquo12) pp 765ndash770 KualaLumpur Malaysia June 2012
[18] S Anwar and L Chen ldquoAn analytical redundancy-based faultdetection and isolation algorithm for a road-wheel controlsubsystem in a steer-by-wire systemrdquo IEEE Transactions onVehicular Technology vol 56 no 5 pp 2859ndash2869 2007
[19] M T Do Z Man C Zhang H Wang and F S Tay ldquoRobustsliding mode-based learning control for steer-by-wire systemsin modern vehiclesrdquo IEEE Transactions on Vehicular Technol-ogy vol 63 no 2 pp 580ndash590 2014
[20] J Duan R Wang and Y Yu ldquoResearch on control strategies ofsteer-by-wire systemrdquo in Proceedings of the International Con-ference on Intelligent Computation Technology and Automation(ICICTA rsquo10) vol 2 pp 1122ndash1125 May 2010
International Journal of Vehicular Technology 17
[21] B Zheng and S Anwar ldquoYaw stability control of a steer-by-wireequipped vehicle via active front wheel steeringrdquoMechatronicsvol 19 no 6 pp 799ndash804 2009
[22] S Sahoo S C Subramanian and S Srivastava ldquoDesign andimplementation of a controller for navigating an autonomousground vehiclerdquo in Proceedings of the 2nd International Confer-ence on Power Control and Embedded Systems (ICPCES rsquo12) pp1ndash6 IEEE Allahabad India December 2012
[23] T D Gillespie Fundamentals of Vehicle Dynamics Society ofAutomotive Engineers 1992
[24] M H Lee S K Ha J Y Choi and K S Yoon ldquoImprovement ofthe steering feel of an electric power steering system by torquemap modificationrdquo Journal of Mechanical Science and Technol-ogy vol 19 no 3 pp 792ndash801 2005
[25] R Pastorino M A Naya J A Perez and J Cuadrado ldquoGearedPM coreless motor modelling for driverrsquos force feedback insteer-by-wire systemsrdquo Mechatronics vol 21 no 6 pp 1043ndash1054 2011
[26] P Koehn andM Eckrich ldquoActive steeringmdashthe BMWapproachtowards modern steering technologyrdquo SAE Technical PaperSeries 2004-01-1105 SAE International 2004
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
International Journal of Vehicular Technology 13
0 2 4 6 8 10minus50
0
50
Time (s)
Rate
chan
ge m
otor
pos
ition
(deg
s)
MeasurementSimulation
(a) Rate change of motor position
0 2 4 6 8 100
10
20
30
40
50
60
Time (s)
Mot
or p
ositi
on (d
eg)
Measurement
Simulation
(b) Position of the motor
0 2 4 6 8 10minus01
minus005
0
005
01
015
Mot
or cu
rren
t (A
)
MeasurementSimulation
Time (s)
(c) Current of the motor
Figure 17 Motor model validation comparison between simulation and measurement
With filter Without filter
Time (s)
Curr
ent (
A)
008
006
004
002
0
minus002
minus004
minus006
minus0081 15 2 25 3 35 4 45 5 55
Figure 18 Noise filtering using low pass filter
in Figure 20(a) From the results the experimental steeringtorque response is almost identical to the EPS steering systemHowever it has small different torque amplitudes These
occur due to the fact that mechanism or friction from themechanical gearbox motor is not modelled in proposedcontrol algorithm
14 International Journal of Vehicular Technology
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
minus152 3 4 5 6 7 8 9 10
(a) Without compensation torque
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
21 3 4 5 6 7 8 9 10
(b) With compensation torque
Figure 19 Steering torque influence on compensation torque
minus80
minus60
minus40
minus20
0
20
40
60
80
100
Stee
ring
whe
el an
gle (
deg)
Time (s)3 4 5 6 7 8 9 10
(a) Steering wheel angle at 80 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
15
20
5
0
minus5
minus10
minus15
minus203 4 5 6 7 8 9 10
(b) Steering torque estimation at 80 kmh
Figure 20 Steering torque estimation and steering wheel angle response at 80 kmh
45 High Speed Manoeuvre At high speed manoeuvre theinput steering wheel angle that is more than plusmn45∘ is very dan-gerous [12] Therefore in this experiment the input steeringwheel angle is less than plusmn45∘ as shown in Figure 21(a) Basedon the results shown in Figures 21(b) and 21(c) the experi-mental steering torque is obtained fairly similar to the EPSsteering system Although the steering wheel angle is at smallangles the influence of vehicle speed cause the steering torqueis increased Moreover at higher speed the vehicle has moreunder steer propensity [9] This increased the self aligningtorque in proportional high torque at front axle motorThusthe driver will sense stiffer on the steering wheel duringmanoeuvre
In general the centering of the steering wheel perfor-mance or steering wheel returnability is better when thesteering torque gradient is large and the hysteresis is small [9]Based on this fact the result is shown in Figure 21(c) whereby
the hysteresis of the steering wheel returnability based onexperimental result is smaller compared to EPS steeringsystem In addition the steering torque gradient is larger thanEPS steering torque Therefore it is considered that the pro-posed feedback torque algorithmnot only achieved a steeringfeel but also improved the steering wheel returnability
46 Lower Speed Manoeuvre The steering torque results atlow speed manoeuvre are shown in Figures 22(b) and 22(c)when the input steering wheel is as shown in Figure 22(a)The vehicle cruised at constant speed of 10 kmh
The experimental steering torque is almost comparableto EPS torque but the amplitude is decreases The torque isdecreased due to the fact that the vehicle speed is reducedwhich is effect on reducing the demand to the total feedbacktorque Moreover the large turning of the steering wheelincreases the backlash of the motor which then contributes
International Journal of Vehicular Technology 15
2 3 4 5 6 7 8 9 10minus30
minus20
minus10
0
10
20
30
40
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle at 120 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steeringSimulationExperiment
10
5
0
minus5
minus10
3 4 5 6 7 8 9
(b) Steering torque estimation at 120 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment10
15
5
0
minus5
minus10
minus15minus20 minus10 0 10 20 30 40
Steering wheel angle (deg)
(c) Steering torque estimation against steering wheel angle at 120 kmh
Figure 21 Steering torque estimation response at 120 kmh
to the decrease of the motor torque Thus driver will sensesofter during steering the wheel during maneuver Howeverit has small different torque against EPS steering Wide rangeadjustment of gains in proposed control algorithm may helpimprove the steering torque response
3 4 5 6 7 8 9 10minus150
minus100
minus50
0
50
100
150
200
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
10
15
5
0
minus5
minus10
minus15
Time (s)2 3 4 5 6 7 8 9 10
(b) Steering torque estimation at 10 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
Steering wheel angle (deg)
10
8
6
4
2
0
minus2
minus4
minus6
minus8
minus10
minus100 minus50 0 50 100 150
(c) Steering torque estimation against steering wheel angle at 10 kmh
Figure 22 Steering torque estimation response at 10 kmh
5 Conclusion and Future Work
The proposed force feedback torque estimation and controlalgorithm have been described The proposed control algo-rithm is considered as an effort to generate the feedback
16 International Journal of Vehicular Technology
torque for driver steering feel in the vehicle SBW systemThe mathematical model of SBW subsystem consists ofsteering wheel and front axle systems that are modelledThe effectiveness of the control algorithm is verified throughthe experimental hardware in the loop (HIL) interfacedwith Matlab XPC target toolbox software According to theexperimental results the proposed control algorithm is ableto generate the feedback torque and the driver is able to sensesteering feel similar to that obtained in EPS steering systemThe LQR controller with gain scheduling based on steeringwheel angle and vehicle speed function provides better torquecontrol that allows rejection of the uncertainty torque froma road condition Furthermore the steering feel gain is ableto increase and lower the steering torque with improvedsteering stability at high speed and it improved the vehiclemanoeuvre at low speed It is also found that compensationtorque is able to improve the feedback torque and to stabilizethe system However it has small different torque magnitudethat occurs due to the backlash of steering wheel motor anduncertainty disturbance of mechanical properties Howeverwith a wide range adjustment of the gains on the proposedcontrol algorithm it may improve and provide realisticsteering torque for better driver steering feel
In future work the proposed control algorithm will betested in real vehicle in order to have more realistic steeringtorque for further investigation and analysis
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
Acknowledgments
This work is fully supported by the Ministry of Higher Edu-cation Malaysia and University Technology Malaysia underFRGS (vote 4F370) and Research University Grant (vote00G64) This work is also supported by Proton HoldingBerhad
References
[1] Y Yao ldquoVehicle steer-by-wire system controlrdquo SAE TechnicalPaper 2006-01-1175 2006
[2] E A Kumar D Dinesh and N Kamble ldquoAn overview of activefront steering systemrdquo International Journal of Scientific ampEngineering Research vol 3 no 6 2012
[3] S M H Fahami H Zamzuri S A Mazlan and N B Zulkar-nain ldquoThe design of vehicle active front steering based on steerby wire systemrdquo Advanced Science Letters vol 19 no 1 pp 61ndash65 2013
[4] S M H Fahami H Zamzuri S A Mazlan and S A SaruchildquoThe variable steering ratio for vehicle steer by wire systemusing hyperbolic tangent methodrdquo Applied Mechanics andMaterials vol 575 pp 781ndash784 2014
[5] T Kaufmann S Milsap B Murray and J Petrowski ldquoDevelop-ment experience with steer by wirerdquo in Proceedings of the SAEInternational Congress and Exhibition SAE 2001-01-2479 2001
[6] S Amberkare F Bolourchi J Demerly and S Millsap ldquoAcontrol system methodology for steer by wire systemsrdquo SAETechnical Paper 2004-01-1106 2004
[7] A Baviskar J R Wagner D M Dawson D Braganza andP Setlur ldquoAn adjustable steer-by-wire haptic-interface trackingcontroller for ground vehiclesrdquo IEEE Transactions on VehicularTechnology vol 58 no 2 pp 546ndash554 2009
[8] M Segawa S Kimura T Kada and S Nakona ldquoA study of reac-tive torque control for steer bywire systemrdquo inProceedings of theInternational Symposium on Advanced Vehicle Control (AVECrsquo02) pp 653ndash658 Hiroshima Japan September 2002
[9] S W Oh H C Chae S C Yun and C S Han ldquoThe design ofa controller for the steer-by-wire systemrdquo JSME InternationalJournal Series C Mechanical Systems Machine Elements andManufacturing vol 47 no 3 pp 896ndash907 2004
[10] C J Kim J H Jang S K Oh J Y Lee and J K HedrickldquoDevelopment of a control algorithm for a rack-actuating steer-by-wire system using road information feedbackrdquo Proceedingsof the Institution of Mechanical Engineers Part D Journal ofAutomobile Engineering vol 222 no 9 pp 1559ndash1571 2008
[11] S Asai H Kuroyanagi and S Takeuchi ldquoDevelopment of asteer-by-wire system with force feedback using a disturbanceobserverrdquo inProceedings of the SAEWorldCongressampExhibitionon Steering amp Suspension Technology Detroit Mich USA 2004
[12] T J Park C S Han and S H Lee ldquoDevelopment of the elec-tronic control unit for the rack-actuating steer-by-wire usingthe hardware-in-the-loop simulation systemrdquo Mechatronicsvol 15 no 8 pp 899ndash918 2005
[13] D Odenhtal T Bunte H D Heitzer and C Eicker ldquoHow tomake steer-by-wire feel like power steeringrdquo in Proceedings ofthe 15th IFACWorld Congress Barcelona Spain July 2000
[14] H Na C Zong and D Hu ldquoInvestigations on corneringcontrol algorithm design and road feeling optimization for asteer-by-wire vehiclerdquo in Proceedings of the IEEE InternationalConference on Mechatronics and Automation (ICMA rsquo09) pp3246ndash3251 Changchun China August 2009
[15] S Ancha A Baviskar J RWagner andDMDawson ldquoGroundvehicle steering systems modelling control and analysis ofhydraulic electric and steer-by-wire configurationsrdquo Interna-tional Journal of Vehicle Design vol 44 no 1-2 pp 188ndash2082007
[16] P Yih and J C Gerdes ldquoModification of vehicle handlingcharacteristics via steer-by-wirerdquo IEEE Transactions on ControlSystems Technology vol 13 no 6 pp 965ndash976 2005
[17] S M H Fahami H Zamzuri S A Mazlan and M A ZakarialdquoModeling and simulation of vehicle steer by wire systemrdquoin Proceedings of the IEEE Symposium on Humanities Scienceand Engineering Research (SHUSER rsquo12) pp 765ndash770 KualaLumpur Malaysia June 2012
[18] S Anwar and L Chen ldquoAn analytical redundancy-based faultdetection and isolation algorithm for a road-wheel controlsubsystem in a steer-by-wire systemrdquo IEEE Transactions onVehicular Technology vol 56 no 5 pp 2859ndash2869 2007
[19] M T Do Z Man C Zhang H Wang and F S Tay ldquoRobustsliding mode-based learning control for steer-by-wire systemsin modern vehiclesrdquo IEEE Transactions on Vehicular Technol-ogy vol 63 no 2 pp 580ndash590 2014
[20] J Duan R Wang and Y Yu ldquoResearch on control strategies ofsteer-by-wire systemrdquo in Proceedings of the International Con-ference on Intelligent Computation Technology and Automation(ICICTA rsquo10) vol 2 pp 1122ndash1125 May 2010
International Journal of Vehicular Technology 17
[21] B Zheng and S Anwar ldquoYaw stability control of a steer-by-wireequipped vehicle via active front wheel steeringrdquoMechatronicsvol 19 no 6 pp 799ndash804 2009
[22] S Sahoo S C Subramanian and S Srivastava ldquoDesign andimplementation of a controller for navigating an autonomousground vehiclerdquo in Proceedings of the 2nd International Confer-ence on Power Control and Embedded Systems (ICPCES rsquo12) pp1ndash6 IEEE Allahabad India December 2012
[23] T D Gillespie Fundamentals of Vehicle Dynamics Society ofAutomotive Engineers 1992
[24] M H Lee S K Ha J Y Choi and K S Yoon ldquoImprovement ofthe steering feel of an electric power steering system by torquemap modificationrdquo Journal of Mechanical Science and Technol-ogy vol 19 no 3 pp 792ndash801 2005
[25] R Pastorino M A Naya J A Perez and J Cuadrado ldquoGearedPM coreless motor modelling for driverrsquos force feedback insteer-by-wire systemsrdquo Mechatronics vol 21 no 6 pp 1043ndash1054 2011
[26] P Koehn andM Eckrich ldquoActive steeringmdashthe BMWapproachtowards modern steering technologyrdquo SAE Technical PaperSeries 2004-01-1105 SAE International 2004
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
14 International Journal of Vehicular Technology
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
minus152 3 4 5 6 7 8 9 10
(a) Without compensation torque
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
5
0
minus5
minus10
21 3 4 5 6 7 8 9 10
(b) With compensation torque
Figure 19 Steering torque influence on compensation torque
minus80
minus60
minus40
minus20
0
20
40
60
80
100
Stee
ring
whe
el an
gle (
deg)
Time (s)3 4 5 6 7 8 9 10
(a) Steering wheel angle at 80 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steering
Simulation
Experiment
10
15
20
5
0
minus5
minus10
minus15
minus203 4 5 6 7 8 9 10
(b) Steering torque estimation at 80 kmh
Figure 20 Steering torque estimation and steering wheel angle response at 80 kmh
45 High Speed Manoeuvre At high speed manoeuvre theinput steering wheel angle that is more than plusmn45∘ is very dan-gerous [12] Therefore in this experiment the input steeringwheel angle is less than plusmn45∘ as shown in Figure 21(a) Basedon the results shown in Figures 21(b) and 21(c) the experi-mental steering torque is obtained fairly similar to the EPSsteering system Although the steering wheel angle is at smallangles the influence of vehicle speed cause the steering torqueis increased Moreover at higher speed the vehicle has moreunder steer propensity [9] This increased the self aligningtorque in proportional high torque at front axle motorThusthe driver will sense stiffer on the steering wheel duringmanoeuvre
In general the centering of the steering wheel perfor-mance or steering wheel returnability is better when thesteering torque gradient is large and the hysteresis is small [9]Based on this fact the result is shown in Figure 21(c) whereby
the hysteresis of the steering wheel returnability based onexperimental result is smaller compared to EPS steeringsystem In addition the steering torque gradient is larger thanEPS steering torque Therefore it is considered that the pro-posed feedback torque algorithmnot only achieved a steeringfeel but also improved the steering wheel returnability
46 Lower Speed Manoeuvre The steering torque results atlow speed manoeuvre are shown in Figures 22(b) and 22(c)when the input steering wheel is as shown in Figure 22(a)The vehicle cruised at constant speed of 10 kmh
The experimental steering torque is almost comparableto EPS torque but the amplitude is decreases The torque isdecreased due to the fact that the vehicle speed is reducedwhich is effect on reducing the demand to the total feedbacktorque Moreover the large turning of the steering wheelincreases the backlash of the motor which then contributes
International Journal of Vehicular Technology 15
2 3 4 5 6 7 8 9 10minus30
minus20
minus10
0
10
20
30
40
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle at 120 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steeringSimulationExperiment
10
5
0
minus5
minus10
3 4 5 6 7 8 9
(b) Steering torque estimation at 120 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment10
15
5
0
minus5
minus10
minus15minus20 minus10 0 10 20 30 40
Steering wheel angle (deg)
(c) Steering torque estimation against steering wheel angle at 120 kmh
Figure 21 Steering torque estimation response at 120 kmh
to the decrease of the motor torque Thus driver will sensesofter during steering the wheel during maneuver Howeverit has small different torque against EPS steering Wide rangeadjustment of gains in proposed control algorithm may helpimprove the steering torque response
3 4 5 6 7 8 9 10minus150
minus100
minus50
0
50
100
150
200
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
10
15
5
0
minus5
minus10
minus15
Time (s)2 3 4 5 6 7 8 9 10
(b) Steering torque estimation at 10 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
Steering wheel angle (deg)
10
8
6
4
2
0
minus2
minus4
minus6
minus8
minus10
minus100 minus50 0 50 100 150
(c) Steering torque estimation against steering wheel angle at 10 kmh
Figure 22 Steering torque estimation response at 10 kmh
5 Conclusion and Future Work
The proposed force feedback torque estimation and controlalgorithm have been described The proposed control algo-rithm is considered as an effort to generate the feedback
16 International Journal of Vehicular Technology
torque for driver steering feel in the vehicle SBW systemThe mathematical model of SBW subsystem consists ofsteering wheel and front axle systems that are modelledThe effectiveness of the control algorithm is verified throughthe experimental hardware in the loop (HIL) interfacedwith Matlab XPC target toolbox software According to theexperimental results the proposed control algorithm is ableto generate the feedback torque and the driver is able to sensesteering feel similar to that obtained in EPS steering systemThe LQR controller with gain scheduling based on steeringwheel angle and vehicle speed function provides better torquecontrol that allows rejection of the uncertainty torque froma road condition Furthermore the steering feel gain is ableto increase and lower the steering torque with improvedsteering stability at high speed and it improved the vehiclemanoeuvre at low speed It is also found that compensationtorque is able to improve the feedback torque and to stabilizethe system However it has small different torque magnitudethat occurs due to the backlash of steering wheel motor anduncertainty disturbance of mechanical properties Howeverwith a wide range adjustment of the gains on the proposedcontrol algorithm it may improve and provide realisticsteering torque for better driver steering feel
In future work the proposed control algorithm will betested in real vehicle in order to have more realistic steeringtorque for further investigation and analysis
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
Acknowledgments
This work is fully supported by the Ministry of Higher Edu-cation Malaysia and University Technology Malaysia underFRGS (vote 4F370) and Research University Grant (vote00G64) This work is also supported by Proton HoldingBerhad
References
[1] Y Yao ldquoVehicle steer-by-wire system controlrdquo SAE TechnicalPaper 2006-01-1175 2006
[2] E A Kumar D Dinesh and N Kamble ldquoAn overview of activefront steering systemrdquo International Journal of Scientific ampEngineering Research vol 3 no 6 2012
[3] S M H Fahami H Zamzuri S A Mazlan and N B Zulkar-nain ldquoThe design of vehicle active front steering based on steerby wire systemrdquo Advanced Science Letters vol 19 no 1 pp 61ndash65 2013
[4] S M H Fahami H Zamzuri S A Mazlan and S A SaruchildquoThe variable steering ratio for vehicle steer by wire systemusing hyperbolic tangent methodrdquo Applied Mechanics andMaterials vol 575 pp 781ndash784 2014
[5] T Kaufmann S Milsap B Murray and J Petrowski ldquoDevelop-ment experience with steer by wirerdquo in Proceedings of the SAEInternational Congress and Exhibition SAE 2001-01-2479 2001
[6] S Amberkare F Bolourchi J Demerly and S Millsap ldquoAcontrol system methodology for steer by wire systemsrdquo SAETechnical Paper 2004-01-1106 2004
[7] A Baviskar J R Wagner D M Dawson D Braganza andP Setlur ldquoAn adjustable steer-by-wire haptic-interface trackingcontroller for ground vehiclesrdquo IEEE Transactions on VehicularTechnology vol 58 no 2 pp 546ndash554 2009
[8] M Segawa S Kimura T Kada and S Nakona ldquoA study of reac-tive torque control for steer bywire systemrdquo inProceedings of theInternational Symposium on Advanced Vehicle Control (AVECrsquo02) pp 653ndash658 Hiroshima Japan September 2002
[9] S W Oh H C Chae S C Yun and C S Han ldquoThe design ofa controller for the steer-by-wire systemrdquo JSME InternationalJournal Series C Mechanical Systems Machine Elements andManufacturing vol 47 no 3 pp 896ndash907 2004
[10] C J Kim J H Jang S K Oh J Y Lee and J K HedrickldquoDevelopment of a control algorithm for a rack-actuating steer-by-wire system using road information feedbackrdquo Proceedingsof the Institution of Mechanical Engineers Part D Journal ofAutomobile Engineering vol 222 no 9 pp 1559ndash1571 2008
[11] S Asai H Kuroyanagi and S Takeuchi ldquoDevelopment of asteer-by-wire system with force feedback using a disturbanceobserverrdquo inProceedings of the SAEWorldCongressampExhibitionon Steering amp Suspension Technology Detroit Mich USA 2004
[12] T J Park C S Han and S H Lee ldquoDevelopment of the elec-tronic control unit for the rack-actuating steer-by-wire usingthe hardware-in-the-loop simulation systemrdquo Mechatronicsvol 15 no 8 pp 899ndash918 2005
[13] D Odenhtal T Bunte H D Heitzer and C Eicker ldquoHow tomake steer-by-wire feel like power steeringrdquo in Proceedings ofthe 15th IFACWorld Congress Barcelona Spain July 2000
[14] H Na C Zong and D Hu ldquoInvestigations on corneringcontrol algorithm design and road feeling optimization for asteer-by-wire vehiclerdquo in Proceedings of the IEEE InternationalConference on Mechatronics and Automation (ICMA rsquo09) pp3246ndash3251 Changchun China August 2009
[15] S Ancha A Baviskar J RWagner andDMDawson ldquoGroundvehicle steering systems modelling control and analysis ofhydraulic electric and steer-by-wire configurationsrdquo Interna-tional Journal of Vehicle Design vol 44 no 1-2 pp 188ndash2082007
[16] P Yih and J C Gerdes ldquoModification of vehicle handlingcharacteristics via steer-by-wirerdquo IEEE Transactions on ControlSystems Technology vol 13 no 6 pp 965ndash976 2005
[17] S M H Fahami H Zamzuri S A Mazlan and M A ZakarialdquoModeling and simulation of vehicle steer by wire systemrdquoin Proceedings of the IEEE Symposium on Humanities Scienceand Engineering Research (SHUSER rsquo12) pp 765ndash770 KualaLumpur Malaysia June 2012
[18] S Anwar and L Chen ldquoAn analytical redundancy-based faultdetection and isolation algorithm for a road-wheel controlsubsystem in a steer-by-wire systemrdquo IEEE Transactions onVehicular Technology vol 56 no 5 pp 2859ndash2869 2007
[19] M T Do Z Man C Zhang H Wang and F S Tay ldquoRobustsliding mode-based learning control for steer-by-wire systemsin modern vehiclesrdquo IEEE Transactions on Vehicular Technol-ogy vol 63 no 2 pp 580ndash590 2014
[20] J Duan R Wang and Y Yu ldquoResearch on control strategies ofsteer-by-wire systemrdquo in Proceedings of the International Con-ference on Intelligent Computation Technology and Automation(ICICTA rsquo10) vol 2 pp 1122ndash1125 May 2010
International Journal of Vehicular Technology 17
[21] B Zheng and S Anwar ldquoYaw stability control of a steer-by-wireequipped vehicle via active front wheel steeringrdquoMechatronicsvol 19 no 6 pp 799ndash804 2009
[22] S Sahoo S C Subramanian and S Srivastava ldquoDesign andimplementation of a controller for navigating an autonomousground vehiclerdquo in Proceedings of the 2nd International Confer-ence on Power Control and Embedded Systems (ICPCES rsquo12) pp1ndash6 IEEE Allahabad India December 2012
[23] T D Gillespie Fundamentals of Vehicle Dynamics Society ofAutomotive Engineers 1992
[24] M H Lee S K Ha J Y Choi and K S Yoon ldquoImprovement ofthe steering feel of an electric power steering system by torquemap modificationrdquo Journal of Mechanical Science and Technol-ogy vol 19 no 3 pp 792ndash801 2005
[25] R Pastorino M A Naya J A Perez and J Cuadrado ldquoGearedPM coreless motor modelling for driverrsquos force feedback insteer-by-wire systemsrdquo Mechatronics vol 21 no 6 pp 1043ndash1054 2011
[26] P Koehn andM Eckrich ldquoActive steeringmdashthe BMWapproachtowards modern steering technologyrdquo SAE Technical PaperSeries 2004-01-1105 SAE International 2004
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
International Journal of Vehicular Technology 15
2 3 4 5 6 7 8 9 10minus30
minus20
minus10
0
10
20
30
40
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle at 120 kmh
Stee
ring
torq
ue (N
m)
Time (s)
EPS steeringSimulationExperiment
10
5
0
minus5
minus10
3 4 5 6 7 8 9
(b) Steering torque estimation at 120 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment10
15
5
0
minus5
minus10
minus15minus20 minus10 0 10 20 30 40
Steering wheel angle (deg)
(c) Steering torque estimation against steering wheel angle at 120 kmh
Figure 21 Steering torque estimation response at 120 kmh
to the decrease of the motor torque Thus driver will sensesofter during steering the wheel during maneuver Howeverit has small different torque against EPS steering Wide rangeadjustment of gains in proposed control algorithm may helpimprove the steering torque response
3 4 5 6 7 8 9 10minus150
minus100
minus50
0
50
100
150
200
Time (s)
Stee
ring
whe
el an
gle (
deg)
(a) Steering wheel angle
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
10
15
5
0
minus5
minus10
minus15
Time (s)2 3 4 5 6 7 8 9 10
(b) Steering torque estimation at 10 kmh
Stee
ring
torq
ue (N
m)
EPS steering
SimulationExperiment
Steering wheel angle (deg)
10
8
6
4
2
0
minus2
minus4
minus6
minus8
minus10
minus100 minus50 0 50 100 150
(c) Steering torque estimation against steering wheel angle at 10 kmh
Figure 22 Steering torque estimation response at 10 kmh
5 Conclusion and Future Work
The proposed force feedback torque estimation and controlalgorithm have been described The proposed control algo-rithm is considered as an effort to generate the feedback
16 International Journal of Vehicular Technology
torque for driver steering feel in the vehicle SBW systemThe mathematical model of SBW subsystem consists ofsteering wheel and front axle systems that are modelledThe effectiveness of the control algorithm is verified throughthe experimental hardware in the loop (HIL) interfacedwith Matlab XPC target toolbox software According to theexperimental results the proposed control algorithm is ableto generate the feedback torque and the driver is able to sensesteering feel similar to that obtained in EPS steering systemThe LQR controller with gain scheduling based on steeringwheel angle and vehicle speed function provides better torquecontrol that allows rejection of the uncertainty torque froma road condition Furthermore the steering feel gain is ableto increase and lower the steering torque with improvedsteering stability at high speed and it improved the vehiclemanoeuvre at low speed It is also found that compensationtorque is able to improve the feedback torque and to stabilizethe system However it has small different torque magnitudethat occurs due to the backlash of steering wheel motor anduncertainty disturbance of mechanical properties Howeverwith a wide range adjustment of the gains on the proposedcontrol algorithm it may improve and provide realisticsteering torque for better driver steering feel
In future work the proposed control algorithm will betested in real vehicle in order to have more realistic steeringtorque for further investigation and analysis
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
Acknowledgments
This work is fully supported by the Ministry of Higher Edu-cation Malaysia and University Technology Malaysia underFRGS (vote 4F370) and Research University Grant (vote00G64) This work is also supported by Proton HoldingBerhad
References
[1] Y Yao ldquoVehicle steer-by-wire system controlrdquo SAE TechnicalPaper 2006-01-1175 2006
[2] E A Kumar D Dinesh and N Kamble ldquoAn overview of activefront steering systemrdquo International Journal of Scientific ampEngineering Research vol 3 no 6 2012
[3] S M H Fahami H Zamzuri S A Mazlan and N B Zulkar-nain ldquoThe design of vehicle active front steering based on steerby wire systemrdquo Advanced Science Letters vol 19 no 1 pp 61ndash65 2013
[4] S M H Fahami H Zamzuri S A Mazlan and S A SaruchildquoThe variable steering ratio for vehicle steer by wire systemusing hyperbolic tangent methodrdquo Applied Mechanics andMaterials vol 575 pp 781ndash784 2014
[5] T Kaufmann S Milsap B Murray and J Petrowski ldquoDevelop-ment experience with steer by wirerdquo in Proceedings of the SAEInternational Congress and Exhibition SAE 2001-01-2479 2001
[6] S Amberkare F Bolourchi J Demerly and S Millsap ldquoAcontrol system methodology for steer by wire systemsrdquo SAETechnical Paper 2004-01-1106 2004
[7] A Baviskar J R Wagner D M Dawson D Braganza andP Setlur ldquoAn adjustable steer-by-wire haptic-interface trackingcontroller for ground vehiclesrdquo IEEE Transactions on VehicularTechnology vol 58 no 2 pp 546ndash554 2009
[8] M Segawa S Kimura T Kada and S Nakona ldquoA study of reac-tive torque control for steer bywire systemrdquo inProceedings of theInternational Symposium on Advanced Vehicle Control (AVECrsquo02) pp 653ndash658 Hiroshima Japan September 2002
[9] S W Oh H C Chae S C Yun and C S Han ldquoThe design ofa controller for the steer-by-wire systemrdquo JSME InternationalJournal Series C Mechanical Systems Machine Elements andManufacturing vol 47 no 3 pp 896ndash907 2004
[10] C J Kim J H Jang S K Oh J Y Lee and J K HedrickldquoDevelopment of a control algorithm for a rack-actuating steer-by-wire system using road information feedbackrdquo Proceedingsof the Institution of Mechanical Engineers Part D Journal ofAutomobile Engineering vol 222 no 9 pp 1559ndash1571 2008
[11] S Asai H Kuroyanagi and S Takeuchi ldquoDevelopment of asteer-by-wire system with force feedback using a disturbanceobserverrdquo inProceedings of the SAEWorldCongressampExhibitionon Steering amp Suspension Technology Detroit Mich USA 2004
[12] T J Park C S Han and S H Lee ldquoDevelopment of the elec-tronic control unit for the rack-actuating steer-by-wire usingthe hardware-in-the-loop simulation systemrdquo Mechatronicsvol 15 no 8 pp 899ndash918 2005
[13] D Odenhtal T Bunte H D Heitzer and C Eicker ldquoHow tomake steer-by-wire feel like power steeringrdquo in Proceedings ofthe 15th IFACWorld Congress Barcelona Spain July 2000
[14] H Na C Zong and D Hu ldquoInvestigations on corneringcontrol algorithm design and road feeling optimization for asteer-by-wire vehiclerdquo in Proceedings of the IEEE InternationalConference on Mechatronics and Automation (ICMA rsquo09) pp3246ndash3251 Changchun China August 2009
[15] S Ancha A Baviskar J RWagner andDMDawson ldquoGroundvehicle steering systems modelling control and analysis ofhydraulic electric and steer-by-wire configurationsrdquo Interna-tional Journal of Vehicle Design vol 44 no 1-2 pp 188ndash2082007
[16] P Yih and J C Gerdes ldquoModification of vehicle handlingcharacteristics via steer-by-wirerdquo IEEE Transactions on ControlSystems Technology vol 13 no 6 pp 965ndash976 2005
[17] S M H Fahami H Zamzuri S A Mazlan and M A ZakarialdquoModeling and simulation of vehicle steer by wire systemrdquoin Proceedings of the IEEE Symposium on Humanities Scienceand Engineering Research (SHUSER rsquo12) pp 765ndash770 KualaLumpur Malaysia June 2012
[18] S Anwar and L Chen ldquoAn analytical redundancy-based faultdetection and isolation algorithm for a road-wheel controlsubsystem in a steer-by-wire systemrdquo IEEE Transactions onVehicular Technology vol 56 no 5 pp 2859ndash2869 2007
[19] M T Do Z Man C Zhang H Wang and F S Tay ldquoRobustsliding mode-based learning control for steer-by-wire systemsin modern vehiclesrdquo IEEE Transactions on Vehicular Technol-ogy vol 63 no 2 pp 580ndash590 2014
[20] J Duan R Wang and Y Yu ldquoResearch on control strategies ofsteer-by-wire systemrdquo in Proceedings of the International Con-ference on Intelligent Computation Technology and Automation(ICICTA rsquo10) vol 2 pp 1122ndash1125 May 2010
International Journal of Vehicular Technology 17
[21] B Zheng and S Anwar ldquoYaw stability control of a steer-by-wireequipped vehicle via active front wheel steeringrdquoMechatronicsvol 19 no 6 pp 799ndash804 2009
[22] S Sahoo S C Subramanian and S Srivastava ldquoDesign andimplementation of a controller for navigating an autonomousground vehiclerdquo in Proceedings of the 2nd International Confer-ence on Power Control and Embedded Systems (ICPCES rsquo12) pp1ndash6 IEEE Allahabad India December 2012
[23] T D Gillespie Fundamentals of Vehicle Dynamics Society ofAutomotive Engineers 1992
[24] M H Lee S K Ha J Y Choi and K S Yoon ldquoImprovement ofthe steering feel of an electric power steering system by torquemap modificationrdquo Journal of Mechanical Science and Technol-ogy vol 19 no 3 pp 792ndash801 2005
[25] R Pastorino M A Naya J A Perez and J Cuadrado ldquoGearedPM coreless motor modelling for driverrsquos force feedback insteer-by-wire systemsrdquo Mechatronics vol 21 no 6 pp 1043ndash1054 2011
[26] P Koehn andM Eckrich ldquoActive steeringmdashthe BMWapproachtowards modern steering technologyrdquo SAE Technical PaperSeries 2004-01-1105 SAE International 2004
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
16 International Journal of Vehicular Technology
torque for driver steering feel in the vehicle SBW systemThe mathematical model of SBW subsystem consists ofsteering wheel and front axle systems that are modelledThe effectiveness of the control algorithm is verified throughthe experimental hardware in the loop (HIL) interfacedwith Matlab XPC target toolbox software According to theexperimental results the proposed control algorithm is ableto generate the feedback torque and the driver is able to sensesteering feel similar to that obtained in EPS steering systemThe LQR controller with gain scheduling based on steeringwheel angle and vehicle speed function provides better torquecontrol that allows rejection of the uncertainty torque froma road condition Furthermore the steering feel gain is ableto increase and lower the steering torque with improvedsteering stability at high speed and it improved the vehiclemanoeuvre at low speed It is also found that compensationtorque is able to improve the feedback torque and to stabilizethe system However it has small different torque magnitudethat occurs due to the backlash of steering wheel motor anduncertainty disturbance of mechanical properties Howeverwith a wide range adjustment of the gains on the proposedcontrol algorithm it may improve and provide realisticsteering torque for better driver steering feel
In future work the proposed control algorithm will betested in real vehicle in order to have more realistic steeringtorque for further investigation and analysis
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
Acknowledgments
This work is fully supported by the Ministry of Higher Edu-cation Malaysia and University Technology Malaysia underFRGS (vote 4F370) and Research University Grant (vote00G64) This work is also supported by Proton HoldingBerhad
References
[1] Y Yao ldquoVehicle steer-by-wire system controlrdquo SAE TechnicalPaper 2006-01-1175 2006
[2] E A Kumar D Dinesh and N Kamble ldquoAn overview of activefront steering systemrdquo International Journal of Scientific ampEngineering Research vol 3 no 6 2012
[3] S M H Fahami H Zamzuri S A Mazlan and N B Zulkar-nain ldquoThe design of vehicle active front steering based on steerby wire systemrdquo Advanced Science Letters vol 19 no 1 pp 61ndash65 2013
[4] S M H Fahami H Zamzuri S A Mazlan and S A SaruchildquoThe variable steering ratio for vehicle steer by wire systemusing hyperbolic tangent methodrdquo Applied Mechanics andMaterials vol 575 pp 781ndash784 2014
[5] T Kaufmann S Milsap B Murray and J Petrowski ldquoDevelop-ment experience with steer by wirerdquo in Proceedings of the SAEInternational Congress and Exhibition SAE 2001-01-2479 2001
[6] S Amberkare F Bolourchi J Demerly and S Millsap ldquoAcontrol system methodology for steer by wire systemsrdquo SAETechnical Paper 2004-01-1106 2004
[7] A Baviskar J R Wagner D M Dawson D Braganza andP Setlur ldquoAn adjustable steer-by-wire haptic-interface trackingcontroller for ground vehiclesrdquo IEEE Transactions on VehicularTechnology vol 58 no 2 pp 546ndash554 2009
[8] M Segawa S Kimura T Kada and S Nakona ldquoA study of reac-tive torque control for steer bywire systemrdquo inProceedings of theInternational Symposium on Advanced Vehicle Control (AVECrsquo02) pp 653ndash658 Hiroshima Japan September 2002
[9] S W Oh H C Chae S C Yun and C S Han ldquoThe design ofa controller for the steer-by-wire systemrdquo JSME InternationalJournal Series C Mechanical Systems Machine Elements andManufacturing vol 47 no 3 pp 896ndash907 2004
[10] C J Kim J H Jang S K Oh J Y Lee and J K HedrickldquoDevelopment of a control algorithm for a rack-actuating steer-by-wire system using road information feedbackrdquo Proceedingsof the Institution of Mechanical Engineers Part D Journal ofAutomobile Engineering vol 222 no 9 pp 1559ndash1571 2008
[11] S Asai H Kuroyanagi and S Takeuchi ldquoDevelopment of asteer-by-wire system with force feedback using a disturbanceobserverrdquo inProceedings of the SAEWorldCongressampExhibitionon Steering amp Suspension Technology Detroit Mich USA 2004
[12] T J Park C S Han and S H Lee ldquoDevelopment of the elec-tronic control unit for the rack-actuating steer-by-wire usingthe hardware-in-the-loop simulation systemrdquo Mechatronicsvol 15 no 8 pp 899ndash918 2005
[13] D Odenhtal T Bunte H D Heitzer and C Eicker ldquoHow tomake steer-by-wire feel like power steeringrdquo in Proceedings ofthe 15th IFACWorld Congress Barcelona Spain July 2000
[14] H Na C Zong and D Hu ldquoInvestigations on corneringcontrol algorithm design and road feeling optimization for asteer-by-wire vehiclerdquo in Proceedings of the IEEE InternationalConference on Mechatronics and Automation (ICMA rsquo09) pp3246ndash3251 Changchun China August 2009
[15] S Ancha A Baviskar J RWagner andDMDawson ldquoGroundvehicle steering systems modelling control and analysis ofhydraulic electric and steer-by-wire configurationsrdquo Interna-tional Journal of Vehicle Design vol 44 no 1-2 pp 188ndash2082007
[16] P Yih and J C Gerdes ldquoModification of vehicle handlingcharacteristics via steer-by-wirerdquo IEEE Transactions on ControlSystems Technology vol 13 no 6 pp 965ndash976 2005
[17] S M H Fahami H Zamzuri S A Mazlan and M A ZakarialdquoModeling and simulation of vehicle steer by wire systemrdquoin Proceedings of the IEEE Symposium on Humanities Scienceand Engineering Research (SHUSER rsquo12) pp 765ndash770 KualaLumpur Malaysia June 2012
[18] S Anwar and L Chen ldquoAn analytical redundancy-based faultdetection and isolation algorithm for a road-wheel controlsubsystem in a steer-by-wire systemrdquo IEEE Transactions onVehicular Technology vol 56 no 5 pp 2859ndash2869 2007
[19] M T Do Z Man C Zhang H Wang and F S Tay ldquoRobustsliding mode-based learning control for steer-by-wire systemsin modern vehiclesrdquo IEEE Transactions on Vehicular Technol-ogy vol 63 no 2 pp 580ndash590 2014
[20] J Duan R Wang and Y Yu ldquoResearch on control strategies ofsteer-by-wire systemrdquo in Proceedings of the International Con-ference on Intelligent Computation Technology and Automation(ICICTA rsquo10) vol 2 pp 1122ndash1125 May 2010
International Journal of Vehicular Technology 17
[21] B Zheng and S Anwar ldquoYaw stability control of a steer-by-wireequipped vehicle via active front wheel steeringrdquoMechatronicsvol 19 no 6 pp 799ndash804 2009
[22] S Sahoo S C Subramanian and S Srivastava ldquoDesign andimplementation of a controller for navigating an autonomousground vehiclerdquo in Proceedings of the 2nd International Confer-ence on Power Control and Embedded Systems (ICPCES rsquo12) pp1ndash6 IEEE Allahabad India December 2012
[23] T D Gillespie Fundamentals of Vehicle Dynamics Society ofAutomotive Engineers 1992
[24] M H Lee S K Ha J Y Choi and K S Yoon ldquoImprovement ofthe steering feel of an electric power steering system by torquemap modificationrdquo Journal of Mechanical Science and Technol-ogy vol 19 no 3 pp 792ndash801 2005
[25] R Pastorino M A Naya J A Perez and J Cuadrado ldquoGearedPM coreless motor modelling for driverrsquos force feedback insteer-by-wire systemsrdquo Mechatronics vol 21 no 6 pp 1043ndash1054 2011
[26] P Koehn andM Eckrich ldquoActive steeringmdashthe BMWapproachtowards modern steering technologyrdquo SAE Technical PaperSeries 2004-01-1105 SAE International 2004
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
International Journal of Vehicular Technology 17
[21] B Zheng and S Anwar ldquoYaw stability control of a steer-by-wireequipped vehicle via active front wheel steeringrdquoMechatronicsvol 19 no 6 pp 799ndash804 2009
[22] S Sahoo S C Subramanian and S Srivastava ldquoDesign andimplementation of a controller for navigating an autonomousground vehiclerdquo in Proceedings of the 2nd International Confer-ence on Power Control and Embedded Systems (ICPCES rsquo12) pp1ndash6 IEEE Allahabad India December 2012
[23] T D Gillespie Fundamentals of Vehicle Dynamics Society ofAutomotive Engineers 1992
[24] M H Lee S K Ha J Y Choi and K S Yoon ldquoImprovement ofthe steering feel of an electric power steering system by torquemap modificationrdquo Journal of Mechanical Science and Technol-ogy vol 19 no 3 pp 792ndash801 2005
[25] R Pastorino M A Naya J A Perez and J Cuadrado ldquoGearedPM coreless motor modelling for driverrsquos force feedback insteer-by-wire systemsrdquo Mechatronics vol 21 no 6 pp 1043ndash1054 2011
[26] P Koehn andM Eckrich ldquoActive steeringmdashthe BMWapproachtowards modern steering technologyrdquo SAE Technical PaperSeries 2004-01-1105 SAE International 2004
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of