history of mechatronics.pdf

7/27/2019 history of mechatronics.pdf

1/17

MECHATRONIC SYSTEMS INNOVATIVE PRODUCTS WITH EMBEDDED CONTROL

Rolf Isermann

Institute of Automatic Control Darmstadt University of Technology, Germany

[email protected]

Abstract: Many technical processes and products in the area of mechanical and electricalengineering are showing an increasing integration of mechanics with digital electronicsand information processing. This integration is between the components (hardware)and the information-driven functions (software), resulting in integrated systems calledmechatronic systems. Their development involves nding an optimal balance betweenthe basic mechanical structure, sensor and actuator implementation, automatic informa-tion processing and overall control. Of major importance are the simultaneous designof mechanics and electronics, hardware and software and embedded control functionsresulting in an integrated component or system. This technical progress has a very largeinuence on a multitude of products in the area of mechanical, electrical and electronicengineering and changes the design, for example, of conventional electromechanicalcomponents, machines, vehicles and precision mechanical devices with increasing in-tensity. This contribution summarizes ongoing developments for mechatronic systems,shows design approaches and examples of mechatronic products and considers variousembedded control functions and systems integrity. One eld of ongoing developments,automotive mechatronics, is described in more detail by discussing mechatronic sus-pensions, mechatronic brakes, active steering and roll stabilization systems.Copyright c IFAC 2005

Keywords: mechatronics, component integration, embedded control, actuators,machines, automobiles, diagnosis, fault-tolerance, hardware-in-the-loop simulation,mechatronic suspensions, mechatronic brakes, active steering, roll-stabilization.

1. INTRODUCTION

Integrated mechanical electronicsystems emerge froma suitable combination of mechanics, electronics andcontrol/information processing. Thereby, these eldsinuence each other mutually. First, a shift of func-tions from mechanics to electronics is observed, fol-lowed by the addition of extended and new functions.

Finally, systems are being developed with certain in-telligentor autonomous functions. For these integratedmechanical electronic systems, the term mechatron-ics has been used for several years.

1.1 From mechanical to mechatronic systems

Mechanical systems generate certain motions or trans-fer forces or torques. For an oriented command of,e.g., displacements, velocities or forces,feedforward and feedback control systems have been applied formany decades. The control systems operate eitherwithout auxiliary energy (e.g., y ball governor), or

with electrical, hydraulic or pneumatic auxiliary en-ergy, to manipulate the commanded variables directlyor with a power amplier. Figure 1 summarizes thisdevelopment, beginning with the purely mechanical


2/17

Pure mechanical systems

1920

1935

1955

1975

1985


3/17

- anti-lock brakingsystem (ABS)

- electro-hydraulic brake (EHB)

- activesuspension

- active frontsteering

Mechatronicsystems

Mechatronicmachine

components

Mechatronictrains

Mechatronicautomobiles

Mechatronic power

consuming

machines

Mechatronic power

producing

machines - tilting trains- active boogie- magnetic

levitated

(MAGLEV)trains

Mechatronicmotion

generators

- integratedmulti-axismachine tools

- integratedhydraulic pumps

- brushless DCmotor

- integrated ACdrives

- mechatroniccombustionengines

- integratedelectricalservo drives

- integratedhydraulicservo drives

- integrated pneumaticservo drives

- robots(multi-axis,mobile)

- semi-activehydraulicdampers

- magnetic bearings

- automaticgears

Fig. 4. Examples for mechatronic systems (macro-mechatronics)

ing with adaptive or even learning behavior which canalso be calledintelligent mechatronic systems .The developments up until now can be followedin (Schweitzer, 1992), (Gausemeieret al. , 1995),(Harashima and Tomizuka, 1996), (Isermann, 1996),(Tomizuka, 2000), (VDI 2206, 2004). An insight intogeneral aspects are given editorially in the journals(Mechatronics, 1991), (IEEE/ASME, 1996), the con-ference proceedings of, e.g., (UK Mechatronics Fo-

rum, 1990, 1992, 1994, 1996, 1998, 2000, 2002),(IMES, 1993), (DUIS, 1993), (ICRAM, 1995), (AIM,1999, 2001, 2003), (IFAC, 2000, 2002, 2004), the journal articles by (Hiller, 1995), (Lckel, 1995), andthe books of (Kitaura, 1986), (Bradleyet al. , 1991),(McConaillet al. , 1991), (Heimannet al. , 2001),(Isermann, 2003), (Bishop, 2002).Mechanical systems can be dedicated to a large areaof mechanical engineering. According to their con-struction, they can be subdivided into mechanicalcomponents, machines, vehicles, precision mechani-cal devices and micromechanical components . Fig-ure 4 shows some examples of mechatronic compo-nents, machinery and vehicles. Examples forpreci-sion mechatronic devices are gyros, laser and ink jetprinters, hard disk drives. Mechatronic products in theeld of microelectromechanical systems (MEMS) arepiezoelectric acceleration sensors, micro actuators andmicropumps.

1.2 Functions of mechatronic systems

1.2.1. Distribution of mechanical and electronic func-tions. Mechatronic systems permit many improvedand new functions. In the design of mechatronic sys-tems, theinterplay for the realization of functions in

the mechanical and electronic part is crucial. Someexamples are:

decentralized electrical drives with microcomputer-control (multi-axis systems, automatic gears);

lightweight constructions : damping by electronicfeedback (drive-trains of vehicles, elastic robots,space constructions);

linear overall behavior of nonlinear mechanicsby proper feedback (hydraulic and pneumatic

actuators, valves); operator adaptation through programmable char-acteristics (accelerator pedal, manipulators).

1.2.2. Operating properties. Process behavior adaptedfeedback control enables for example:

increase of mechanical precision by feedback; adaptive friction compensation ; model-based andadaptive control to allow wide-

range operation (ow-, force-, speed-control, en-gines, vehicles, aircraft);

high control performance due to closer set-points to constraints (engines, turbines, papermachines).

1.2.3. New functions. Mechatronic systems makefunctionspossible that could notbe performedwithout(embedded) digital computers, like:

control of nonmeasurable variables (tire slip,internal tensions or temperatures, slip angle andground speed of vehicles, damping parameters);

advanced supervision andfault diagnosis ;

fault-tolerant systems with hardware and analyt-ical redundancy; teleservice functions for monitoring, mainte-

nance, repair;


4/17

Integration by information processing

on-line information processing

feedforward,feedback control

supervisiondiagnosis

adaptationoptimization

knowledge base

informationgaining:-identification-state observer

design methods:-control-supervision-optimization

performancecriteria

mathematical process models

actuator process sensors

Integration of componentsmicro-

computer

Fig. 5. Integration of mechatronic systems: integrationof components (hardware integration); integra-tion by information processing (software integra-tion)

exible adaptation to changing boundary condi-tions;

programmable functions allow changes duringdesign, commissioning and after-sales, and shortertime-to-market.

1.3 Integration forms

With increasing improvements of the miniaturization,robustness and computing power of microelectroniccomponents, one can try to put more weight on theelectronic side and to design the mechanical part fromthe beginning with a view to amechatronic overallsystem . Then, more autonomous systems can be envis-aged, e.g., in the form of capsular units with wirelesssignal transfer or bus connections and robust micro-

electronics. The integration within a mechatronic sys-tem can be performed mainly in two ways, through theintegration of components and through integration byinformation processing.Theintegration of components (hardware integration )results from designing the mechatronic system as anoverall system and embedding the sensors, actuatorsand microcomputers into the mechanical process, seeFigure 5. This spatial integration may be limited to theprocess and sensor or the process and actuator. Themicrocomputers can be integrated with the actuator,the process or sensor, or be arranged at several places.

Integrated sensors and microcomputers lead tosmart sensors and integrated actuators and microcomputersdevelop intosmart actuators . For larger systems, busconnections will replace the many cables.

Integration by information processing (software inte-gration ) is mostly based on advanced control func-tions. Besides a basic feedforward and feedback con-trol, an additional inuence may take place throughthe process knowledge and corresponding on-line in-formation processing in higher levels, see Figure 5.This includes the solution of tasks like supervisionwith fault diagnosis, optimization and general processmanagement. The respective problem solutions re-sult in anon-line information processing , especiallyby real-time algorithms, which must be adapted tothe mechanical process properties, e.g., expressed bymathematical models. Therefore, aknowledge base isrequired, comprising methods for design and informa-tion gain, process models and performance criteria. Inthis way, the mechanical parts are governed in vari-ous ways through higher level information processingwith intelligent properties, possibly including learn-ing, thus forming anintegration by process adapted

software .

2. DESIGN PROCEDURE

The design of mechatronic systems requires a sys-tematic development and use of modern software de-sign tools. As with any design, mechatronic design isalso an iterative procedure. However, it is much moreinvolved than for pure mechanical or electrical sys-tems. In addition to the traditional domain specic en-gineering (mechanical, electrical/electronic, automa-tion, user interface) an integrated, simultaneous (con-current) engineering is required. It is theintegrationof engineering across traditional boundaries that istypical for the development of mechatronic systems.A representation of important design steps, which dis-tinguishes especially between themechatronic systemdesign and system integration is depicted in Figure6. This scheme is represented in form of a V-model, which originates probably from software de-velopment, (STARTS GUIDE, 1989), (Brhl, 1995),see also (VDI 2206, 2004).Within this V-model-development scheme only someexamples for specic mechatronic issues are consid-ered here. Thesystem design includes the task dis-tribution between mechanical, hydraulic, pneumatic,electrical and electronic components, the used aux-iliary power, the type and placement of sensors andactuators, the electronic architecture, the software ar-chitecture, the control engineering design and the cre-ation of synergies. Because of the many varieties of designs the advancemodelling and simulation playsan important role, also to save the number of realizedprototypes. Therefore, theoretical/physical modellingof the heterogeneous components is required, using

general modelling principles. For this purpose object-oriented software tools like DYMOLA, MODELICA,MOBILE, VHDL-AMS, 20 SIM are especially suit-able, see (Otter and Cellier, 1996), (Elmqvist, 1993),


5/17

validation

degree of maturity

verification

s y s t e m d e s i g n

- mechanics- electronics

Requirements- overall functions- rated values- costs & milestones

Specifications- fulfillment of requirements- sources, limitations

- reliability, safety

Modeling & Simulation- models of components- behavior analysis- requirements for components

design

System design- partitioning- modules- mechanics vs. electronics- synergies

Prototypes- laboratory solutions- modifications of former products- prototype computers/algorithms

Production- simultaneous planning- technologies- assembling- quality control

Field testing- final product- normal use- statistics- certification

System testing- test rigs- stress testing, EMC- behavior testing- reliability, safety

System integration (software)- signal analysis- filtering- tuning of algorithms

System integration (hardware)- assembling- mutual adaptationoptimization- synergies

Component testing / tuning- hardware-in-the-loop simulation- stress analysis

s y s t e m

i n t e

g r a t i o n

mechanics electronics automaticcontrol

human-machineinterface

- control-software- human-machine interface

Mechatronic components

ECU

BypassComputer

ECU

BypassComputer

ECU

BypassComputer

EngineSimulation

EngineSimulation

EngineSimulation

Component design (domain specific)

Fig. 6. V development scheme for mechatronic systems

(Hiller, 1995), (van Amerongen, 2004), together withsimulation tools like MATLAB/SIMULINK.In this stage of development, use is made of software-in-the-loop simulation (SiL), i.e., components andcontrol algorithms are simulated on an arbitrary com-puter without real-time requirements to obtain, e.g.,design specications, dynamic requirements and per-formance measures. Thecomponent design is domainspecic and uses general CASE-tools, like CAD/CAEfor mechanics, CFD-tools for uidics, circuit boardlayout-tools (PADS), microelectronic design tools(VHDL) and CADCS-tools for automatic control de-sign. Also the reliability and safety is considered, seeSection 4. Then prototypes are build as laboratory so-lutions. The system integration begins with rst stepsto combine the different components. Because of thedifferent development status of the components duringthe simultaneous design, minimization of iterative de-velopment cycles and meeting of short time-to-marketschedules, frequently use is made of different kind of real-time simulations , Figure 7 and (VDI 2206, 2004).A rst case is therapid control prototyping (RCP)where the real process is operated together withthe simulated control by a high-speed hardware andsoftware other than the nal electronic control unit(ECU) (either full-passing or partially by-passing theECU with special software functions on the RCP-computer). A second case is thehardware-in-the-loop simulation (HiL), where the real-time simulatedprocess runs with the real ECU hardware and alsoactuator hardware. This is an especially demanding

task, because the real-time process simulation must berather precise and the sensor outputs signals have tobe realized with special interface circuits. Advantagesof HiL are, e.g., testing in laboratory environment,

testing under extreme operating conditions and withfaults, reproductive experiments, design of human-machine interface. Thesystem integration comprisesthe spatial integration of thehardware components byembedding the sensors, actuators, cables and buseson or into the mechanics and creation of synergeticeffects and the functional integration by thesoftwarewith all algorithms from control through adaptationto supervision, fault diagnosis, fault tolerance and hu-man/machine operation.

3. AUTOMATIC CONTROL OF MECHATRONICSYSTEMS

3.1 Control design

The applied feedforward and feedback control al-gorithms depend on the individual properties of theelectrical, mechanical, hydraulic, pneumatic and alsothermal processes. They can be brought into a gen-eral knowledge-based multi-level control structure asshown in Figure 8. The knowledge base consists of mathematical process models, identication and para-

meter estimation algorithms, controller design meth-ods and control performance criteria. The feedbackcontrol can be organized into lower level and higherlevel controllers, a reference value generation module


6/17

SiL

T

p2

2

p2, stat

T 2, stat

q FW

n eng

controlalgorithm

p2

p2, Setpoint

q FW n eng

u pwm

process model

simulation tool

real process (engine) real ECU + real actuator(injection pump)

high performancereal-time computer(full pass, bypass)

integratedmechatronic

system(final product)

ECU model

H i L R C P

Fig. 7. Various kinds of combining real and simulated parts for development: SiL: Software-in-the-loop; RCP:rapid control prototyping; HiL: Hardware-in-the-loop

controller design

parameter estimation

mathem.models

sensors

k n o w l e d g e b a s e

f e e d b a c k c o n t r o l l e v e l s

perfor-mancecriteria

processactuatorsmanualcontrol

u y,

u

w

w

controller adaptation

referencegeneration

higher levelcontroller

lower levelcontroller

y

y

y22

11

Fig. 8. Knowledge-based multi-level feedback controlfor mechatronic systems

and controller parameter adaptation. Because of thelarge variety of possibilities, only some control princi-ples will be considered briey. More methods are, forexample, presented in (Spong and Vidyasagar, 1989),(Morari and Zarov, 1989), (strm and Wittenmark,1997), (Isidori, 1999), (Dorf and Bishop, 2001) and(Goodwinet al. , 2001).Some basic design requirements are limited compu-tations because of real-time constraints, nonlinearityof the processes, limited actuator speed and range,robustness, transparency of solutions, maintainability,etc. Of major importance is thesimultaneous design of the mechatronic process and thecontrol . This meansthat the static and dynamic behavior of the process, thetype and placement of the actuators, the type and posi-tion of the sensors are designed appropriately, result-ing in an control dynamic friendly overall behavior.(A very important item for any control design.)The goal of thelower level feedback is to provide acertain dynamic behavior (e.g., enforcement of damp-

ing), to compensate nonlinearities like friction, to re-duce parameter sensitivity and to stabilize. Some typ-ical examples are: Damping of high-frequency oscillations : weaklydampedoscillations appear, e.g., in multi-mass drivetrains orpneumatic and hydraulic actuators. The damping cangenerally be improved by high-pass ltering the out-puts and using a state variable or PD (proportional-

derivative) feedback.Compensation of nonlinear static characteristics : non-linear static characteristics are present in many sub-systems of mechanical processes. Figure 9 shows atypical example, the often required position controlfor a nonlinearactuator. Frequently, a rstnonlinearityappears in the force- or torque-generating part like anelectromagnet or a pneumatic or hydraulic actuatorwhere, e.g., the forceF D = f (U ) follows a nonlin-ear static characteristic. This nonlinearity can now becompensated by an inverse characteristicU = f 1(U )such that the I/O behaviorF D = f (U ) becomes ap-

proximately linear and a linear (PID-type) controllerGc1 can be applied.Friction compensation : for many mechanical systems,the overall friction can be described approximately by

F F (t ) = f FC signY (t ) + f Fv Y (t ) |Y (t )| > 0 (1)

wheref FC is the Coulomb friction andf F the linearviscous friction coefcient which may be dependenton the motion direction, indicated by+ or . TheCoulomb friction has a strong negative effect on the

control performance. Different methods such as com-pensating the relay function, see Figure 9, dithering,feedforward compensation and adaptive friction com-pensation are alternatives, see, e.g., (Isermann and


7/17

parameter estimation & controller design & supervision

positioncontroller withfrictioncompensation

casade control non-linear actuator

force controller with non-linear force compen-sation

force generation

c M F c

F F U FC

F L F

D m mGc2 Gc1

Q 1QQ 2Q

r

U Y Y

Y Y

.

.U

mechanical motion

1m

U

Fig. 9. Adaptive position control of a nonlinear electromechanical, hydraulic or pneumatic actuator (example)

Raab, 1993), (Tomizuka, 1995) and (Canudas de Witet al. , 1995).An alternative for position control of nonlinear actua-tors is the use of asliding mode controller . It consistsof a nominal part for feedback linearization and an ad-ditional feedback to compensate for model uncertain-ties, (Utkin, 1977), (Slotine and Weiping, 1991). Theresultingchattering by the included switchingfunctionhereby generates a dither signal. A comparison with axedPID-controller with friction compensation shows(Pfeuferet al. , 1995).

Stabilization : unstable mechatronic systems like mag-netic bearings, magnetic levitated trains or skiddingautomobiles have to be stabilized in the lower controllevel by appropriate feedback laws. The stabilizationfeedback usually includes derivative terms and in thecase of magnetic actuators compensating terms for thenonlinearities.Switching actuator control : low-cost actuators in theform of solenoids or pneumatic membrane-types areusually manipulated by pulse-width-modulated inputsignals of higher frequency allowing approximatelylinear behavior for control of their position or uid

pressure in the lower frequency band.The control scheme of the lower level control maybe expanded by additional feedback controllers froma load or working process that is coupled with themechanical process, resulting in a multiple-cascadedcontrol system. A prerequisite for the application of advanced control algorithms is the use of well-adaptedprocess models. This may lead to self-tuning oradap-tive control systems .The task of thehigher level controller is to generate agood overall dynamic behavior with regard to changesof the position referencer (t ) and to compensate forexternal disturbances stemming, e.g., from load varia-tions. This high-level controller may be realized as aparameter optimized controller of PID-type or internal

model controller or a state controller with or withoutstate observer.Parameter scheduling : parameter (or gain) schedulingbased on the measurement of, e.g., load-dependentvariables is an effective method to deal with knownvarying process behavior.Parameter-adaptive control systems : parameter-adaptivecontrol systems are characterized by using identi-cation methods for parametric process models. Thisis indicated in the adaptation level of Figures 8 and9. Parameter estimation has proven to be an appro-

priate basis for the adaptive control of mechanicalprocesses, including the adaptation to nonlinear char-acteristics, Coulomb friction, and the unknown para-meters like masses, stiffness, damping, see (Isermannand Raab, 1993), (Isermannet al. , 1992), (strm andWittenmark, 1997).If no appropriate sensors to measure the controlledvariable are availablefeedforward control has to beused. Feedforward controls may be realized as simpleproportional or proportional-derivative algorithms, asstatic nonlinear characteristicsu = f ( y) or as nonlinearlook-up tables/maps u = f (y). The last case holds,

e.g., for the control of internal combustion engineswhere low-cost sensors for torque and emissions arenot available and stability problems have to be avoidedunder all circumstances.A considerable part of the automation of mechatronicsystems is performed bysequential control , e.g., forprocesses with repetitive operation (machine-tools,printing machines), start-up and shut-down (engines)or automatic gears. Hence, mechatronic systems makeuse of a large variety of different control methods,ranging from simple proportional or on-off controllersto internal model and adaptive nonlinear controllers.

Because the process model structure is mostly known,structure optimized controllers can be realized. Theprocess model order is usually not large, but nonlin-


8/17

change detection

fault diagnosis

diagnosed faults

signal-basedfault detection process

model-basedfault detection

normal behavior

s analytical symptoms

f

limit-

plausibilitycheck

vibration

signalmodels

process

processmodel

featuregeneration

faults

actuatorsU

N Y

x featuresr ,Q ,

sensors

Fig. 10. Scheme for model-based fault detectionearities and especially the actuator behavior has tobe taken into the design. A big advantage is that theprocess and its control is delivered from one manu-facturer, such that optimal controllers can be imple-mented and maintained, however, subject to consider-able real-time constraints. A detailed description canbe given only for concrete processes, actuation princi-ples and measurement congurations.

3.2 Supervision and fault diagnosis

As the right functioning of mechatronic systems de-pends not only on the process itself, but also onthe electronic and electrical sensors, actuators, cables,plugs and electronic control units, anautomatic supervision (health monitoring) and if possible,fault detection anddiagnosis plays an increasingly impor-tant role, especially with regard to high reliability andsafety requirements.Figure 10 shows a process inuenced by faults. Thesefaults indicate unpermitted deviations from normalstates and are generated either externally or inter-nally. External faults are, e.g., caused by the powersupply, contamination or collision, internal faults bywear, missing lubrication, actuator or sensor faults.The classical methods for fault detection are thelimit value checking

orplausibility checks

of a few mea-surable variables. However, incipient and intermittentfaults usually cannot be detected and an in-depth faultdiagnosis is not possible with this simple approach.Therefore,signal- and process-model-based fault de-tection and diagnosis methods have been developedin recent years, allowing early detection of smallfaults with normally measured signals, also in closedloops, (Isermann, 1997), (Gertler, 1998), (Chen andPatton, 1999). Based on measured input signalsU (t ),output signalsY (t ) and process models, features aregenerated by, e.g.,vibration analysis , parameter esti-mation , state and output observers and parity equa-tions , Figure 10.These features are then compared with the features fornormal behavior, and with change detection methods,

analytical symptoms are obtained. Then, afault di-agnosis is performed via methods of classication orreasoning .A considerable advantage is that the same processmodel can be used for both the (adaptive) controllerdesign and the fault detection. In general, continuous

time models are preferred if fault detection is basedon parameter estimation or parity equations. However,discrete time models can also be used. Advanced su-pervision and fault diagnosis is a basis for improv-ing reliability and safety, state-dependent maintenanceand triggering of redundancies and reconguration forfault-tolerant systems, (Isermann, 2005).

4. RELIABILITY, SAFETY, FAULT TOLERANCE

Compared to pure mechanic, hydraulic or pneumaticsystems, mechatronic systems replace very reliable

mechanical parts by less reliable electrical, and elec-tronic components and software. Therefore, the de-sign must be paralleled byreliability analysis pro-cedures like event tree analysis (ETA), fault-treeanalysis (FTA) and failure mode and effect analy-sis (FMEA), (IEC 60812, 1985), (IEC 61508, 1997),(Storey, 1996), (Onodera, 1997). By using probabilitymeasures like failure rates or MTTF (mean-time-to-failure) it is tried to nd weak spots of the design inearly and later stages of development.For safety-related systems ahazard-analysis with riskclassication has to be performed, e.g., by statingquantitative risk measures based on the probabilityand consequences of dangers and accidents. Safetyintegrity levels (SiL) are introduced for different kindsof processes, like stationary machinery, automobiles,aircraft etc. After applying reliability and safety analy-sis methods during design and testing and qualitycontrol during manufacturing the development of cer-tain faults and failures still cannot be avoided totally.Therefore, especially high-integrity systems require fault-tolerance . This means that faults are compen-sated such that they do not lead to system failures.Fault-tolerance methods generally use redundancy.This means that in addition to the considered mod-ule, one or more modules are connected, usually inparallel. These redundant modules are either identicalor diverse. Such redundant schemes can be designedfor hardware, software, information processing, andmechanical and electrical components like sensors,actuators, microcomputers, buses, power supplies, etc.There exist mainly two basic approaches for fault-tolerance, static redundancy and dynamic redundancy,see Figure 11. Fault tolerance with dynamic redun-dancy and cold standby is especially attractive formechatronic systems where more measured signals

and embedded computers are already available andtherefore fault detection can be improved considerablyby applying process model-based approaches. Follow-ingsteps of degradation are distinguished:


9/17


10/17

Mechatronicautomobiles

Mechatroniccombustion

engines

Mechatronicsteering

Mechatronic brakes

Mechatronicsuspensions

Mechatronicdrivetrains

- parameter-izable power-assistedsteering

- electro-mechanical power-assistedsteering (EPS)

- active frontsteering (AFS)

- hydraulic anti-lock braking(ABS)

- electronicstability program (ESP)

- electro-hydraulic brake (EHB)

- electro-mechanical brake (EMB)

- electrical parking brake

- semi-activeshock-absorbers

- active hydr.suspension(ABC)

- active pneumaticsuspension

- active anti-roll bars(dynamic drivecontrol (DDC)or roll-control)

- automatichydrodynamictransmission

- automaticmechanicshift transm.

- continuouslyvariable trans-mission (CVT)

- automatictraction control(ATC)

- automaticspeed anddistance control(ACC)

- electricalthrottle

- mechatronicfuel injection

- mechatronicvalve trains

- variable geo-metry turbo-charger (VGT)

- emissioncontrol

- evaporativeemissioncontrol

- electrical pumps & fans

Fig. 12. Survey of mechatronic components and systems for automobiles and engines

ringchannel

suspension

controller feedback level"fast"

performanceindex l }

parameter estimation

valve

(b)

computationof

coefficients

systemexcitation

processsignals

verticalmovement

z , z , z , z

c , c , d , m , F

B

W

W W B

B B B C

systemexcitation

r F ,

u

x

B

a , b ii^

^ ^ ^ ^ ^

^

}adap-tationlevel"slow"

(a)

I

Fig. 13. Semi-active shock absorber (a) and its control(b)

Active suspensions provide an extra force input in ad-dition to existing passive springs. They may be real-ized as hydraulic, hydro pneumatic or pneumatic sys-tems. The required energy is for passenger cars and anoperating range between 0 to 5 Hz about 1-2kW andbetween 0-12Hz about 2-7kW. Figure 14 shows asone example a hydraulic active suspension with a hy-draulic piston in series with the steel spring, (Merkeret al. , 2001). This concept is designed to reduce lowfrequent body motions ( f < 2 Hz), due to rolling andpitching and to reduce higher frequent road excita-tions ( f < 6 Hz). It is controlled by a state-feedback

controller with measurement of deectionz BW be-tween body and wheel and body accelerationz B. Arecent survey on mechatronic suspensions is givenby (Fischer and Isermann, 2004) and a model-based

z t

i (t ) , u (t ) , (t )

body

c B

d B

B z

R z

M

motor hydraulic

acculumator

pump linear bearing

plunger

hydraulicline

uM , iM M ,

electro-hyd. susp.

body-controller

M M M

-

ref

(t)

z S

z WB

z P

()

0

()

()WB

B

z t

z t

WB, ref

] ]

] ]

Fig. 14. Active hydraulic suspension system (ABC,Mercedes CL and S-class), measured signals

fault detection of an active suspension by (Fischeret al. , 2004).

5.2 Mechatronic brake systems

The conventional hydraulic brake systems with twoindependent, redundant hydraulic circuits are the stan-dard solution for passenger cars. However, due todriver assisting functions like ABS and ESP they be-come more complex. In order to increase the function-ality further, to safe space and assembling costs and to

increase the passive safety, two types of mechatronicbrake-by-wire systems were developed, the electrohy-draulic brake (EHB), since 2001 in series production(Mercedes SL and E-class), and the electromechanical


11/17

(b)

(a)

piston pump& motor

ECUhydro-module

high pressurestorage

Fig. 15. Illustration of brake-by-wire-systems: (a)Electrohydraulic brake control (EHB), Bosch;(b) Electromechanical brake (EMB), ContinentalTeves

brake (EMB), for which prototypes exist, see Figure15.Figure 16 shows the different stages forbrake systemsof passenger cars or light weight trucks. In the case of theconventional hydraulic brake , the mechanical link-age between the pedal and the hydraulic main cylin-der is paralleled by the power supporting pneumaticactuator (booster). If the pneumatic actuator fails, themechanical linkage transfers the (larger) pedal forcefrom the driver. The hydraulic cylinder acts on twoindependent hydraulic circuits in parallel. That meansthe brake system is fault-tolerant with regard to afailure of one of the two hydraulic circuits. Failuresin the electronics of brake control systems as ABSbring the hydraulic actuators (e.g., magnetic valves)

into a fail-safe status such that the hydraulic brake getsthe pressure from the hydraulic main cylinder directly.The ABS functions are realized by switching valves,which have three positions for lowering, holding orincreasing the uid pressure and thus allow only adiscrete actuation of the brake torque, with strongoscillations.A rst step towards brake-by-wire is theelectrohy-draulic brake (EHB), Figures 16a) and 15a), wherethe mechanical pedal has sensors for position and hy-draulic pressure, (Jonneret al. , 1996), (Stoll, 2001).Their signals are transferred to separated hydraulic

pressure loops with proportional magnetic valves, ma-nipulating hydraulic liquid ows from a 160bar stor-age/pump system to the wheel brakes. If the electron-ics fail the separation of the pedal to the wheel brakes

Fig. 18. Braking with model-based ABS- (anti-lock-braking system) functions, measured on dry as-phalt. Continuous nonlinear, adaptive slip con-trol with EHB (electrohydraulic brake) generatesmaximal brake froces (FL, RL: front, rear left)

is released. Hence, a hydraulic back-up serves to failsafe as for conventional hydraulic brakes.Theelectromechanical brake (EMB) according to Fig-ures 16b) and 15b) does not contain hydraulics any-more. The pedal possesses sensors and its signals aresent to a central brake control computer and wheelbrake controllers which both act through power elec-tronics to the electromotors of, e.g., disc brakes. Be-cause no mechanical or hydraulic connection does ex-ist a mechanical or hydraulic fail-safe is not possible.Hence, the complete electrical path must be build with

fault tolerance, see the architecture in Figure 17. Both,the EHB and EMB, have many advantages with regardto control functions . One important property is theability to continuously manipulate the brake torqueduring ABS actions. Figure 18 shows an example forfull braking with ABS functions based on continu-ous, proportional acting slip controlled EHB-brakes,(Semmleret al. , 2002). The applied controller is afeedback linearized nonlinear controller which opti-mizes the slip to result in maximal braking forces. Ex-cept EHB, the further introduction of brake-by-wire,like EMB, is not decided yet.

However,presently the introduction of complete brake-by-wire and steer-by-wire systems is undecided, be-cause many functions can also be realized withelectromechanical and electrohydraulic systems, toohigh costs and missing 42V-electrical system.

5.3 Mechatronic steering systems

Hydraulic assisted power steering goes back untilaround 1945. This classical steering was continuouslyimproved, especially in adapting the required force

or torque support to the speed. Later developmentsrealized this reducing support with increasing speedby electronically controlled electromagnetic by-passvalves, also called parameterizable steering. Figure


12/17

T

T

V . p p

p I

I

U

U

V

V electro-hydraul.actuator

hydraulic pump/

accumul.

.

.

hydrauliccylind.

cut valves

bus

w

w

hydraulic brake wheel vehicle

brakecontrol

(ABS,...)

wdriver mech./

electron. pedal .

electrical supportingenergy

electricalenergy (42 V)

welectr.mech. brake

power electron.amplif.

central brakecontrol

wheel vehicle

wheel brakecontrol

z,

driver mech. pedal

pedalelectr.

(a)

(b)

j

D

D

D

D

p

p

F F

F

z

T

w

T w

Fig. 16. Signal ow diagram for different mechatronic brake systems of passenger cars: (a) Electrohydraulic brake(EHB) with hydraulic brake; (b) Electromechanical brake (EMB) without mechanical backup

wheel brakecontroller 3

(duplex)

wheel brakemodule 3(fail-safe)

wheel brake

actuator 3

connect.


(duplex)


wheel brake

actuator 4


(duplex)


wheel brake

actuator 1


bus 2

bus 1

vehicle bus

c o n n e c

t .

c o n n e c

t .

switch 2

sensor element 3

sensor element 4

ASIC(duplex 2)

duplex box 2


(duplex)

wheel brake

actuator 2

switch 1

sensor element 1

sensor element 2

ASIC(duplex 1)

duplex box 1

central controler (duplex)

central controller module (fail-silent)

battery 2

battery 1

vehicle electric system

EMB busDirect SignalsEMB ElectricPower Supply

brake pedal module(fail-operational)

pedal sensors andelectronics:duo-duplex

connect. connect. connect.

connect. connect.

Fig. 17. Fault-tolerant electromechanical brake (EMB) system architecture (prototype)

HydraulicPower Steering

(HPS)

ElectricalPower Steering

(EPS)

ActiveFront Steering

(AFS)

Steer-by-Wire

(SbW)

(e)

Electrical Power Assisted Steering

(HPS+EPS)

(d)(c)(b)(a)

Fig. 19. Mechatronic steering systems: (a) conventional hydraulic power steering (HPS) (since about 1945); (b)electrical power steering (EPS) for smaller cars (1996); (c) electrical power assisted steering (HPS+EPS) forlarger cars; (d) active front steering (AFS): Additional wheel angles generated by a planetary gear and a DCmotor (2003); (e) steer-by-wire (SbW). (Not introduced by now.)


13/17

hapticactuator.

electr.power 1

electr.power 2

low energy level

high energy level

mechan.pedal/wheel

T T

I

U

hapticfeedback

brake/steer

control

manage-ment

supervis.

sensors

sensorspedal/wheel

electronics

actuator

control

bus-

system

power electron-

icsvehicle

electr.

actuator

micro-computer

mechan./electronicpedal

brake/steer controlsystem

brake/steer

mechan.w wdriver

T w

A

A

W

W

D

D

Fig. 20. Signal ow diagram of drive-by-wire systems

19 shows some mechatronic steering systems. Sinceabout 1996 electrically assisted power steering (EPS)is on the market for smaller cars, (Connor, 1996).For larger cars the hydraulic power steering (HPS) isparalleled by electrical power steering (EPS), allow-ing electrical inputs for, e.g., automatic parking. Arecent development is theactive front steering (AFS)introduced in 2003, where additional steering anglesare generated with a DC motor acting on a planetarygear, (Koniket al. , 2000). By this construction themechanical linkage to the wheels is maintained andelectrical inputs can be superimposed. This enables toincrease the steering gain with lower speed, a higherdynamic steering and allows yaw changes and, e.g.,sidewind compensation.Figure 20, shows a general signal ow diagram of a drive-by-wire system . The drivers operating unit(steering wheel, braking pedal) has a mechanical input(e.g., torque or force) and an electrical output (e.g.bus protocol). It contains sensors and switches forposition and/or force, microelectronics and either apassive (spring-damper) or active (el. actuator) feed-back to give the driver a haptic information ("pedal-feeling") on the action. A bus connects to the brakesor steer control system with actuator control, brakeor steer function control, supervision and differentkinds of management (e.g. fault tolerance with recon-guration), (Stlzl, 2000), (Stlzlet al. , 1998) and

(Isermannet al. , 2002). Important are redundant elec-trical power supplies (12 V and 42V), like two batter-ies and a generator.

5.4 Active front steering control with active anti-rollbar stiffness variation (an example)

An active front steering according to Figure 19d)is considered which generates steering angles c(t )through a planetary gear and a DC motor in additionto the drivers steering angle (t ), Figure 21, see e.g.,

(Ackermannet al. , 1995). The task of the steeringfeedback controller shown in Figure 22 is to manip-ulate the steering angle sum w(t ) such that the yawrate behavior( ) for the driver is close to a one-track

Fig. 21. Active front steering, generating additionalsteering angles through a planetary gear andbrushless DC motor (BMW).

reference model. This reference model is identicalwith the expected steering behavior for normal drivingsituations with velocity-dependent behavior and in-cludes the dynamics of lateral tire forces, (Schornet al. , 2005). The steering controller is a PID-controllerwith velocity-dependent parameters designed with lo-cal linearone-track models. Figure 23 shows thesimu-lated resulting steering behavior for a vehicle enteringa curve with 144km/h using a veried comprehensivetwo-track model including lateral and vertical dynam-ics. The one-track model yields in the case of station-ary cornering for the required steering angle relatedto the Ackermann steering angle 0 = il l/

0

= 1+1

v2ch

v2 = 1+ SGl

v2 (2)

with

SG =l

v2ch(under) steer gradient (3)

v2ch =c F c R l2

m(c R l R c F lF )characteristic velocity (4)

(c wheel cornering stiffness,lF , R distance betweenCG and axle,F front,R rear,l = lF + l R).Thus,(SG > 0),(SG = 0),(SG < 0) indicates under-steering, neutral steering and oversteering.After entering the curve att = 0 s, the vehicle startsto oversteer att = 2s because the velocity is by far


14/17


15/17

M stab

F stab,r

F stab,r

active anti-roll bar

c ( )stiff,f c stiff,r ( )

=1

=0

v

Rear Front

a)

b)

Fig. 24. Active anti-roll bars: a) schematic diagram;b) distribution parameter of the anti-roll barstiffness

Fig. 25. Active anti-roll bar system for roll stabiliza-tion and steering support (BMW)

include complex control algorithms, condition moni-toring and fault-diagnosis methods and require fault-tolerant components for safety-related processes.The contribution gives an overview of the structureand design of mechatronic systems and considers var-ious embedded control functions and fault tolerance.As example for innovations the economically impor-tant area of automotive mechatronics is highlighted.Mechatronic suspensions, brake systems and steeringsystems change the design of automobiles fundamen-tally, improving functionality, safety, economy andcomfort. Similar developments can be observed for

combustion engines, trains, aircraft, machine tools,and automation components, etc. Thus, mechatronicdevelopment is an emerging area for innovative engi-neering.

REFERENCES

Ackermann, J., J. Guldner, W. Sienel, R. Steinhauserand V.I. Utkin (1995). Linear and nonlinear con-troller design for robust automatic steering.IEEE Control System Technology 3, 132140.

AIM (1999, 2001, 2003).IEEE/ASME Conference

on Advanced Intelligent Mechatronics. Atlanta(1999), Como (2001), Kobe (2003) .strm, K.H. and B. Wittenmark (1997).Computer-

controlled Systems. Theory and Design . PrenticeHall. Upper Saddle River.

Bishop, C.M. (2002).The mechatronics handbook .CRC Press. Boca Raton.

Bradley, D.A., D. Dawson, D. Burd and A.J. Loader(1991).Mechatronics-electronics in products and processes . Chapman and Hall. London.

Brhl, A. P., Ed.) (1995).Das V-Modell - Der Stan-dard fr Softwareentwicklung, 2nd edn. . Olden-bourg. Mnchen.

Buhardt, J. and R. Isermann (1993). Parameter adap-tive semi-active shock absorbers. In:ECC Eu-ropean Control Conference . Vol. 4. Groningen,Netherlands. pp. 22542259.

Canudas de Wit, C., H. Olsson, K.J strm and P. Lin-schinsky (1995). A new model for control of sys-tems with friction.IEEE Trans. on AutomaticControl 40 , 419425.

Chen, J. and R.J. Patton (1999).Robust model-based fault diagnosis for dynamic systems . Kluwer.Boston.

Connor, B. (1996). Elektrische Lenkhilfen fr Pkwals Alternative zu hydraulischen und elektrischenSystemen.Automobiltechnische Zeitschrift 98(7-8), 406410.

Dieterle, W. (2004). Mechatronic systems: industrialapplications and modern design methodology. In:3rd IFAC Symposium on Mechatronic Systems .Sydney, Australia.

Dorf, R.C. and R.H. Bishop (2001).Modern controlsystems, 9th ed. . Prentice Hall. Englewood Cliffs.

DUIS (1993).Mechatronics and Robotics. M. Hiller, B. Fink (eds). 2nd Conference, Duisburg/Moers,Sept 27-29 .

Elmqvist, H. (1993).Object-oriented modeling and automatic formula manipulation in Dymola .Scandin. Simul. Society SIMS. Kongsberg.

Fischer, D. and R. Isermann (2004). Mechatronicsemi-active and active vehicle suspensions.Con-trol Engineering Practice 12 , 13531367.

Fischer, D., H.-P. Schner and R. Isermann (2004).Model-based fault detection for an active ve-hicle suspension. In:FISITA World AutomotiveCongress . Barcelona, Spain.

Gausemeier, J., D. Brexel, T. Frank and A. Humpert(1995). Integrated product development. In:3rd

Conference on Mechatronics and Robotics . Teub-ner, Stuttgart. Paderborn, Germany.Gertler, J. (1998).Fault detection and diagnosis in

engineering systems . Marcel Dekker. New York.


16/17

Goodall, R. (1995). Mechatronics in motion - somerailway applications. In:3rd IFAC Symposium on Mechatronic Systems . Sydney, Australia.

Goodwin, G.C., S.F. Graebe and M.E. Salgado (2001).Control system design . Prentice Hall. EnglewoodCliffs.

Guzella, L. and C.H. Onder (2004).Introduction tomodeling and control of internal combustion en-gine systems . Springer. Berlin.

Harashima, F. and M. Tomizuka (1996). Mechatronics what it is, why and how?.IEEE/ASME Trans.on Mechatronics 1, 12.

Heimann, B., W. Gerth and K. Popp (2001).Mecha-tronik . Fachbuchverlag Leipzig. Leipzig.

Hiller, M. (1995). Modelling, simulation and controldesign for large and heavy manipulators. In:In-ternational Conference on Recent Advances in Mechatronics . Istanbul, Turkey.

ICRAM (1995).Recent Advances in Mechatron-ics. Proceedings of International Conference ICRAM95, Istanbul, August 14 16 .

IEC 60812 (1985).Analysis techniques for systemreliability procedure for failure mode and effectsanalysis (FMEA) . International ElectrotechnicalCommission. Switzerland.

IEC 61508 (1997).Functional safety of elec-trical/electronic/programmable electronic sys-tems . International Electrotechnical Commission.Switzerland.

IEEE/ASME (1996).Transactions on Mechatronics .Vol. 1.

IFAC (2000, 2002, 2004).IFAC-Symposium on Mechatronic Systems: Darmstadt (2000), Berke-ley (2002), Sydney (2004) . Elsevier. Oxford.

IFAC-T.C 4.2. (2000). IFAC Technical Committee on Mechatronics Systems.http://rumi.newcastle.edu.au/reza/TCM/ .

IMES (1993).Integrated Mechanical Electronic Sys-tems Conference (in German) TU Darmstadt, March 2-3 . Vol. Fortschr.-Ber. VDI Reihe 12.VDI-Verlag. Dsseldorf.

Isermann, R. (1996). Modeling and design methodol-ogy of mechatronic systems.IEEE/ASME Trans.on Mechatronics 1, 1628.

Isermann, R. (1997). Supervision, fault-detection andfault-diagnosis methods. an introduction.Control Engineering Practice 5(5), 639652.

Isermann, R. (2000). Mechatronic systems: conceptsand applications.Trans. of the Institute of Mea-surement and Control .

Isermann, R. (2003).Mechatronic Systems. (Germanedition: 1999) . Springer. Berlin.

Isermann, R. (2005).Fault diagnosis and fault toler-ance . Springer. Heidelberg, Berlin.

Isermann, R. and U. Raab (1993). Intelligent actuators- ways to autonomous actuating systems.Auto-

matica 29(5), 13151331.Isermann, R., K.-H. Lachmann and D. Matko (1992). Adaptive Control Systems . Prentice Hall Interna-tional UK. London.

Isermann, R., R. Schwarz and S. Stlzl (2002). Fault-tolerant drive-by-wire systems.IEEE ControlSystems Magazine (October), 6481.

Isidori, A., Ed.) (1999).Nonlinear Control Systems II .Springer. London.

Johansson, R. and Rantzer, A., Eds.) (2003).Nonlin-ear and hybrid systems in automotive control .Springer. London.

Jonner, W.D., H. Winner, L. Dreilich and E. Schunck(1996). Electrohydraulic brake system - the rstapproach. In:SAE Technical paper Series . num-ber 960991. Warrendale.

Kiencke, U. and Nielsen, L., Eds.) (2000).Automotivecontrol systems. For engine, driveline and vehi-cle . Springer. Berlin.

Kitaura, K. (1986).Industrial Mechatronics (in Japanese) . New East Business Ltd.

Konik, D., R. Bartz, F. Brnthol, H. Brunds andM. Wimmer (2000). Dynamic drive - the new ac-tive roll stabilization system from bmw group.In:Proceedings of AVEC 2000, 5th InternationalSymposium on Advanced Vehicle Control .

Kyura, N. and H. Oho (1996). Mechatronics anindustrial perspective..IEEE/ASME Trans. on Mechatronics 1, 1015.

Lckel, J., Ed.) (1995).Third Conference on Mecha-tronics and Robotics. Paderborn, Oct. 4-6 . Teub-ner. Stuttgart.

McConaill, P.A., Drews, P. and Robrock, K.-H., Eds.)(1991).Mechatronics and robotics . ICS Press.Amsterdam.

Mechatronics (1991).An International Journal. Aimsand Scope . Pergamon Press. Oxford.

Merker,T., J. Wirtz, M. Hiller andM. Jeglitzka (2001).Das SL-Fahrwerk..ATZ - Automatisierungstech-nische Zeitschrift. Special Issue: Der neue Mer-cedes SL pp. 8491.

Morari, M. and F. Zarov (1989).Robust processcontrol . Prentice Hall. Englewood Cliffs.

Onodera, K. (1997). Effective techniques of fmea ateach life-cycle stage. In:1997 Proceedings: An-nual Reliability and Maintainability Symposium .IEEE. pp. 5056.

Otter, M. and C. Cellier (1996). Software for modelingand simulating control systems. In:The Control Handbook (W.S. Levine, Ed.). CRC Press. BocaRaton.

ttgen, O. and T. Bertram (2002). Inuencing vehi-cle handling through active roll moment distribu-tion. In:Proceedings of AVEC 2002. 6th Interna-tional Symposium on Advanced Vehicle Control .Hiroshima, Japan. pp. 129134.

Pfeufer, T., T. Landsiedel and R. Isermann (1995).Identication and model-based nonlinear con-trol of electro-mechanical actuators with friction.

In:IFAC-Workshop Motion Control

. Munic, Ger-many.Schmitt, J., R. Isermann, M. Brner and D. Fischer

(2005). Model-based supervision and control of


17/17

lateral vehicle dynamics.Control EngineeringPractice .

Schner, H.P. (2004). Automotive mechatronics.Con-trol Engineering Practice 12 , 13431351.

Schorn, M., J. Schmitt, U. Sthlin and R. Isermann(2005). Model based braking control with sup-port by active steering. In:Proceedings of the16th IFAC World Congress 2005, Prague, Czech Republic .

Schweitzer, G. (1992). Mechatronics a concept withexamples in active magnetic bearings.Mecha-tronics 2, 6574.

Semmler, S., R. Isermann, R. Schwarz and P. Rieth(2002). Wheel slip control for antilock-brakingsystems using brake-by-wire actuators. Vol. SAE2002-01-0303. Detroit, USA.

Slotine, J.J.E. and L. Weiping (1991).Applied nonlin-ear control, Chapter 7 in Sliding Control . Pren-tice Hall. Englewood Cliffs.

Spong, M.W. and M. Vidyasagar (1989).Robust dy-namics and control . J. Wiley. New York.STARTS GUIDE (1989).A guide to methods and

software tools for the construction of largereal-time systems. . National Computing Centre.Manchester.

Stoll, U. (2001).Sensotronic brake control (SBC) . Vol.VDI-Ber. 1646. VDI. Dsseldorf.

Stlzl, S. (2000).Das elektrohydraulische Bremssys-tem von Contintental Teves . Vol. Fortschr.-Ber.VDI Reihe 12. VDI-Verlag. Dsseldorf.

Stlzl, S., R. Schwarz, R. Isermann, J Bhm, J. Nelland P. Rieth (1998). Control and supervision of an electromechanical brake system. In:FISITAWorld Automotive Congress, The Second Centuryof the Automobile . Paris, France.

Storey, N. (1996).Safety-critical computer systems .Addison Wesely Longman Ltd.. Essex.

Tomizuka, M. (1995). Robust digital motioncontrollerformechanical systems. In: International Confer-ence on Recent Advances in Mechatronics . Istan-bul, Turkey.

Tomizuka, M. (2000). Mechatronics: from the 20thto the 21th century. In:1st IFAC Conference on Mechatronic Systems . Elsevier, Oxford. Darm-

stadt, Germany.UK Mechatronics Forum (1990,1992, 1994, 1996, 1998, 2000, 2002).Confer-ences in Cambridge (1990), Dundee (1992), Bu-dapest (1994), Guimaraes (1996), Skovde (1998), Atlanta (2000), Twente (2002) .

Utkin, V.I. (1977). Variable structure systems withsliding mode: a survey.IEEE Trans. on Auto-matic Control 22 , 212222.

van Amerongen, J. (2004). Mechatronic educationand research 15 years of experience. In:3rd IFAC Symposium on Mechatronic Systems . Syn-dey, Australia. pp. 595607.

VDI 2206 (2004).Entwicklungsmethodik fr mecha-tronische Systeme . Beuth Verlag. Berlin.

Date post:	02-Apr-2018
Category:	Documents
Upload:	jaya-seelan-ceg
View:	219 times
Download:	0 times

history of mechatronics.pdf

Documents