Research ArticleA Novel Linear Active Disturbance Rejection Control Design forAir-Breathing Supersonic Vehicle Attitude System withPrescribed Performance
Chao Ming and Xiaoming Wang
School of Mechanical Engineering, Nanjing University of Science and Technology, Nanjing 210094, China
This paper investigates the design problem of the attitude controller for air-breathing supersonic vehicle subject to uncertaintiesand disturbances. Firstly, the longitudinal model is established for the attitude controller design which is devised as a strictfeedback formulation, and a transformed tracking error is derived with the prescribed performance control technique such thatit can limit the tracking error to a predefined region. Then, a novel linear active disturbance rejection control scheme isproposed for the attitude system to enhance the steady-state and transient-state performances by incorporating the transformedtracking error. On the basis of the Lyapunov stability theorem, the convergence and stability characteristics are both rigorouslyproved for the closed-loop system. Finally, extensive contrast simulations are conducted to demonstrate the effectiveness,robustness, and advantage of the proposed control strategy.
1. Introduction
Hypersonic technique is a wide development prospect inboth military and civil fields [1, 2]. As a derivative of hyper-sonic technique, air-breathing supersonic vehicle (ASV) is anew type of aircraft with large airspace, super flight velocity,long range, and high precision compared with the traditionalaircraft [3], and flight guidance-control design is a key tech-nology in the ASV system. However, as the special structureand changeable flight conditions, the vehicle dynamics arepeculiar including the fast time varying, nonlinearity, uncer-tainty, and multiple disturbances [4], which lead to more dif-ficulties and complexities in the control design and analysis.In addition, the ASV usually attacks the target at a super-sonic velocity such that the flight states are changed rapidlywhich imposes a higher requirement for the transient char-acteristic of the control system [5]. Therefore, the controlperformance index such as the overshoot, steady-state error,convergence rate, and robustness must be considered themajor designed indexes in the control system design of suchflight vehicle.
In the past few years, the air-breathing vehicle techniquehas attracted the wide attention of domestic and foreign sci-entific research institutions and scholars due to the afore-mentioned superiorities, and various control methods havebeen explored for such vehicles, such as linear parametervarying (LPV) method [6], dynamic inversion [7, 8], trajec-tory linearization control [9, 10], fuzzy control method[11], dynamic surface control [12, 13], back-stepping controlmethod [14, 15], neural network [16, 17], and sliding modecontrol [18–21].
It is worth noting that the above listed control methodachieved an outstanding control effect for air-breathing vehi-cle, but most mainly fasten the attention on the stability,robustness, and accuracy of the control system. However,the situation of autodisturbance rejection is seldom consid-ered in control design. As we all know that the active distur-bance rejection control (ADRC) method can achieve asatisfactory performance for nonlinear systems, where theparameter perturbations and external disturbances can beestimated and rejected actively, it is successfully applied inindustrial control [22, 23]. However, there is an inevitable
HindawiInternational Journal of Aerospace EngineeringVolume 2020, Article ID 1676739, 15 pageshttps://doi.org/10.1155/2020/1676739
problem that many parameters are needed to be tuned whichlimits the application of ADRC in practice. In this connec-tion, a linear active disturbance rejection control (LADRC)was firstly designed by Prof. Gao [24] which is extremely sim-ple and easily implementable and it has been widely extendedto various industrial control fields, such as electric erectionsystem [25], hovercraft system [26], servo systems [27], windpower systems [28], and photobioreactor [29]. Meanwhile,[30, 31] adopt the LADRC approach to design the attitudecontrol for a supersonic missile. As far as we know, thereare few literatures which concern with the attitude controlsystem design by adopting the LADRC method. Further-more, in the design process of the above LADRC scheme,both did not consider the transient performance design.Therefore, it is vitally essential for the ASV attitude controlsystem design that can guarantee the system for both thesteady-state performance and transient performance withLADRC method.
Recently, a newly emerging control method called theprescribed performance control (PPC) for the nonlinear sys-tem was firstly proposed by Na et al. [32], where the trackingerror can converge to an arbitrarily predefined small set andthe convergence rate and maximum overshoot can be delim-ited less than a prespecified constant. Owing to the significantadvantages in improving the control performance, the PPC isintroduced to the vehicle suspension system [33], manipula-tor system [34], servo mechanisms [35, 36], unmanned aerialvehicle [37], spacecraft [38–40], etc., and several new ideas ofthe PPC design are emerged; [41] proposes a new error trans-formation method and the performance function so that thelimitation of PPC on the known initial error can be relaxed.[42] investigates a new PPC methodology for the longitudi-nal dynamic model of an air-breathing hypersonic vehiclevia neural approximation that the satisfactory transient per-formance with small overshoot can be achieved. [43] presentsa new error transformation to reduce the complexity of thecontrol system which is caused by conventional error con-straint approaches. Berger [44] proposes a novel prescribedperformance controller which can guarantee the output tostay within a prescribed performance funnel bound by incor-porating the funnel control with the PPC technique. [45]designs a new prescribed performance controller withoutan approximation structure which can avoid the largeamount of calculation and some specific problems of thefuzzy or neural network method in the approximation pro-cess. Although extensive achievements have been yielded intheory and application of the PPC technique, the previouscontroller design methods are mainly neural network con-trol, backstepping control, dynamic surface control, and slid-ing model control. To our best knowledge, no results on theLADRC design with the PPC technique have been reportedfor nonlinear systems.
Motivated by the aforementioned discussions, this studyinvestigates a novel LADRC scheme for the longitudinal atti-tude mode of the air-breathing supersonic vehicle in thepresence of uncertain dynamics and external disturbanceswith a prescribed performance constraint. The main contri-butions of this paper are summarized as follows: (1) A novelLADRC approach design method with a prescribed perfor-
mance constraint is firstly proposed for the ASV attitude sys-tem with multiple disturbances. (2) The proposed controlscheme does not need the knowledge of the flight dynamicmodel, and the uncertainty and disturbance can be activelyestimated and compensated into the control signal. (3) Theproposed controller can improve the steady-state and tran-sient performances of the ASV attitude system comparedwith the traditional LADRC method. (4) The system stabilityand convergence characteristic are both proved strictly.
The rest of this paper is outlined as follows. The consid-ered ASV attitude control model is established and the priorknowledge and problem formulation is given in Section 2.Section 3 presents the LADRC-based prescribed perfor-mance attitude controller design procedures. Comparativesimulation results are provided in Section 4. Finally, someconclusions of this work are presented in Section 5.
2. Problem Statement and Preliminaries
In the section, the considered ASV attitude dynamic model ispresented; then, some basic definitions of the prescribed per-formance control method are provided for the subsequentanalysis, and finally, the objective of controller design isoutlined.
2.1. Vehicle Attitude Model. For the subsequent design, thenonlinear longitudinal model of the ASV attitude system inthis paper is derived on the basis of the model which includesthe altitude, velocity, and attitude motion of the hypersonicvehicle modeled in [46, 47]. In addition, the main purposeis to investigate a novel attitude control method of ASV,and the attitude dynamics belongs to a fast motion comparedwith the change of velocity and altitude. With design simpli-fication and no loss of generality, the changes of altitude andvelocity can be ignored due to the limited influence on atti-tude dynamics. Hence, the ASV attitude system can be sim-plified as follows:
_ϑ = ωz ,
_ωz =Mz
Jz,
α = ϑ − θ,
8>>>><>>>>:
ð1Þ
where the pitch moment Mz is defined as
Mz = �qSL m0 +mαzα +mα2
z α2� �
+ �qSLmδzz δz , ð2Þ
where elevator deflection δz is the control input, and m0,mαz ,
mα2z , andmδz
z are the aerodynamic coefficients.For simplicity, defining the state ½x1, x2� = ½ϑ, ωz� and
combining with (1) and (2), the ASV attitude control systemcan be described with the strict feedback state-space form:
_x1 = x2,
_x2 = f + bu,
(ð3Þ
2 International Journal of Aerospace Engineering
where
f =�qSL m0 +mα
zα +mα2z α2
� �Jz
+ Δ
b =�qSLmδz
z
Jz
u = δz
8>>>>>>><>>>>>>>:
ð4Þ
Here, the term f is described as “total disturbance,”which consists of the unknown dynamic uncertainty andexternal disturbance.
Remark 1. The uncertainty term Δ is mainly the resultof parameter uncertainties and external disturbances,e.g., the aerodynamic parameters perturbations and windinterference.
Assumption 1. The total disturbance f is differentiable, andits derivative _f is described as _f = hðx1, x2, ΔÞ . The distur-bance Δ in system (3) is unknown, but the disturbance andits derivative are all bounded.
2.2. Prescribed Performance Theory. The prescribed perfor-mance denotes that the prescribed transient and steady-state performances can strictly limit the tracking error to apredesigned small residual set. Based on the prescribed per-formance concepts [31], the prescribed performance can beobtained in the condition that the tracking error eðtÞ evolvesin the predefined bounds with a decreasing smooth functionas follows:
−ρ tð Þ < e tð Þ < ρ tð Þ, ∀t > 0: ð5Þ
The prescribed performance function (PPF) ρðtÞ isselected as [38]
ρ tð Þ = ρ0 − ρ∞ð Þ exp −κtð Þ + ρ∞, ð6Þ
where ρ0 > ρ∞ > 0 and κ > 0. The maximum overshoot,steady-state error, and convergence rate will be limited byselecting the appropriate parameters ρ0, ρ∞, and κ, respec-tively. For a more intuitive insight of the concept, Figure 1illustrates the schematic diagram of the aforementioned pre-scribed performance theory.
It is worth to mention that the complexity of the control-ler design will be significantly increased by directly adopting(5) and (6) which correspond to an additional constraint ofthe controlled system. In order to evade this problem, the sys-tem with constrains can be converted to an equivalent disen-gaged one by introducing an error transformation functionTð⋅Þ, i.e.,
e tð Þ = ρ tð ÞT ε tð Þð Þ, ð7Þ
where εðtÞ is the transformed tracking error, and Tð⋅Þ issmooth and strictly increasing which satisfies the conditions−1 < TðεÞ < 1 and lim
ε→+∞TðεÞ = 1, lim
ε→−∞TðεÞ = −1.
Here, we design the error transformation function Tð⋅Þ as
where λ = eðtÞ/ρðtÞ is the normalized tracking error.
Remark 2. In the subsequent parts, the transformed error εðtÞwill be used into the controller in place of the tracking erroreðtÞ to deal with the problem of control system design withprescribed performance constraint.
2.3. Control Objective. The control objective is the proposednovel controller u for the ASV attitude system (3) that thestate x1 track the desired command x1c accurately, and thetracking error e1 = x1c − x1 can be limited within a predefinedbound with a satisfactory prescribed performance in spite ofmultiple disturbances including unmodeled dynamics,uncertainties, and external disturbances, i.e., the objectivecan be expressed as follows:
(1) The state x1 can accurately track the desired com-mand x1c with the unknown multiple disturbances
(2) The output tracking error e1 is stabilized at the originwith a prescribed maximum overshoot, the steady-state error, and convergence rate
(3) The closed-loop system states are both stable androbust to the uncertainties and disturbances
3. Control System Design and Stability Analysis
In this part, a novel ASV attitude controller is proposed basedon the LADRC method by introducing the PPC technique,which can improve the steady-state and transient perfor-mances of the ASV attitude control system. Then, the
e (0)
0
e (t)
t (s)
𝜌 (t)
–𝜌 (t)
–𝜌0
𝜌0
–𝜌∞
𝜌∞
Figure 1: The prescribed performance illustration.
3International Journal of Aerospace Engineering
convergence property and stability analysis for the controlsystem are provided on the basic of the Lyapunov method.
3.1. Attitude Control System Design. The block diagram of thepresented control system is depicted in Figure 2, whichmainly consists of three parts: prescribed performance con-trol (PPC), linear extended state observer (LESO), and linearstate error feedback (LSEF).
In view of the previous description, the design process ofthe ASV attitude control system is illustrated as follows:
Step 1.The LESO is the core part of the LADRC; it can generate
the estimation of the states and the disturbances in real time;the estimated value can be used to compensate the distur-bances to the controller which can enhance the robustnessof the system. According to system (3), a three-order LESOis constructed as follows:
where z1, z2, and z3 are the observer states, and β1, β2, and β3are the designed gains which can be chosen with the pole-placement method [24] as follows:
β1 = 3ω0,
β2 = 3ω20,
β3 = ω30,
ð11Þ
where ω0 is the observer bandwidth.With the properly selected gains, z1, z2 will converge to
the system states x1 and x2, respectively, and z3 will accu-rately track the total disturbance term f , i.e., z1 ⟶ x1,z2 ⟶ x2, z3 ⟶ f .
Note that the LESO can estimate the states and totaldisturbance exactly without the mathematical precisemodel, it is only dependent of the system input and outputinformation.
Step 2.The PPC can transform the tracking error into an equiv-
alent unconstrained one by incorporating the transformation
function and the prescribed performance function such thatthe tracking error can be limited in the envelope of the pre-scribed performance bounds (PPB).
Here, we define the error ~e1 as
~e1 = x1c − z1, ð12Þ
where x1c is the desired command and z1 is the estimationvalue of x1 obtained by LESO.
Subsequently, the normalized error λ1 is given by
λ1 =~e1
ρ1 tð Þ , ð13Þ
where ρ1ðtÞ is a similar PPF defined in (6), and it is expressedas
ρ1 tð Þ = ρ10 − ρ1∞ð Þ exp −κ1tð Þ + ρ1∞, ð14Þ
whereρ10,ρ1∞, and κ1 are all positive constants.Then, the transformed error is obtained by
Step 3.Based on the accurate estimation and the designed trans-
formed error, the LSEF can approximately simplify the sys-tem to a disturbance-free form meanwhile reconciling theperformance of the system. The final control law is designas follows:
u = u0 + uf , ð16Þ
where term u0 denotes a linear state error feedback controlterm that guarantees the system is asymptotically stable,and term uf represents the dynamic compensation controlterm to suppress the unfavourable consequence of the totaldisturbance such that it can enhance the robustness of theASV attitude control system.
Here, the control subitem u0 adopted the linear pro-portional and derivative (PD) control framework by
Errornormalized
Errortransformed
PPC
LSEF
Totaldisturbances
PPF
PPC-based LADRC scheme
LESO
ASVattitudedynamic
e1 𝜆1 𝜀1 𝛿zx1c
z1z2 z3
𝜌1 (t)
Figure 2: The block diagram of the presented controller.
4 International Journal of Aerospace Engineering
introducing the transformed error ε1 of PPC which isdesigned as follows:
u0 =kpε1 − kdz2
b, ð17Þ
where kp, kd are the controller gains and z2 is the estima-tion value of x2 obtained by LESO, and expression of b isgiven in (3).
Meanwhile, the control subitem uf compensates the dis-turbances with the estimated value z3, which is given by
uf = −z3b: ð18Þ
Thus, the controller u is obtained in terms of (16), (17),and (18) as follows:
u =kpε1 − kdz2 − z3
b: ð19Þ
Note that the controller u0 introduces the transformed errorε1 instead of ~e1 which can enhance the transient and steady-state performances with the PPC technique.
3.2. Stability Analysis. In this subsection, the observer con-vergence and the closed-loop system stability will be analysedwith the listed theorem. Prior to investigating, the followinglemmas are introduced.
Lemma 1. For system _ηðtÞ =NηðtÞ + gðtÞ, where N is a n × nmatrix, and ηðtÞ = ½η1ðtÞ, η2ðtÞ,⋯, ηnðtÞ�, gðtÞ = ½g1ðtÞ,g2ðtÞ,⋯, gnðtÞ�, if lim
t→∞kgðtÞk = 0 and N is Hurwitz, then
limt→∞
kηðtÞk = 0 holds.
This lemma has been proofed in detail which can be seenin [48].
Lemma 2. If the transformed error εðtÞ can be controlled to bebounded, i.e., the condition jεðtÞj ≤ εM holds where εM > 0 isconstant; then, the tracking error eðtÞ can be controlled withina prescribed boundary, i.e., −ρðtÞ < eðtÞ < ρðtÞ, ∀t > 0 holds.
This lemma has been proofed in detail which can be seenin [38].
Theorem 1. For system (3) and LESO (10), if _f = hðx1, x2, ΔÞsatisfies the Lipschitz condition in the definition domain, thereexist a constant ω0 > 0 such that the estimated states z1, z2,and z3 can converge to the state x1, x2, and f , respectively,i.e., the observer errors satisfy that lim
t→∞ez iðtÞ = 0, i = 1, 2, 3,
where the observer errors are defined as ez1 = x1 − z1, ez2 = x2− z2, and ez3 = f − z3.
Proof. Define the observer errors as
ez1 = x1 − z1,
ez2 = x2 − z2,
ez3 = f − z3:
ð20Þ
From (3) and (10), the observer error dynamic can begiven by
_ez1 = −3ω0 ez1 + ez2,
_ez2 = −3ω20 ez1 + ez3,
_ez3 = −ω30 ez1 + _f :
8>><>>: ð21Þ
Let ~e = ½~e1,~e2,~e3� = ½ez1, ez 2/ω0, ez 3/ω02� and _f = hðx1,
x2, ΔÞ, then (21) can be rewritten as follows:
_~e = ω0~A~e + ~B h x1, x2, Δð Þ
ω03 , ð22Þ
where ~A =
−3 1 0
−3 0 0
−1 0 0
2664
3775, ~B =
0
0
1
2664
3775.
As ~Asatisfies the Hurwitz stability, the condition P~A +~ATP = −I holds, where the matrix I is 3-order identity matrixand P is a positive definite Hermitian matrix and isexpressed as
where k > 0 is constant.Meanwhile, as the equation −2~e1 −~e2 + 8~e3 = 2~eT P~B
holds, we can obtain that
−2~e1 −~e2 + 8~e3ð Þ h x1, x2, Δð Þj jω0
2 ≤ 2~eTP~Bk ~ek kω0
2 : ð28Þ
In addition, the following inequality holds
P~Bk�� ��2 − 2P~Bk + 1 ≥ 0: ð29Þ
Thus, we obtain
2~eTP~Bk ~ek kω0
2 ≤ K ~ek k2, ð30Þ
where K = ðkPBZkk2 + 1Þ/ω02.
Based on (26), (28), and (30), we can obtain that
_V1 ≤ − ω0 − Kð Þ ~eZk k2: ð31Þ
That is, if ω0 > K such that _V1 < 0. Therefore, limt→∞
ez iðtÞ =0, i = 1, 2, 3 can be achieved, i.e., the estimation errors of theLESO (10) are asymptotically stable. The proof of Theorem1 is completed.
Theorem 2. For system (3), LESO (10), and the proposedcontroller (19) with the prescribed performance function(14), if _f = hðx1, x2, ΔÞ satisfies the Lipschitz condition indefinition domain, there exists constants ω0 > 0 and ωc > 0that can guarantee the system (3) to be asymptotically stable.Furthermore, the tracking error e1 can be maintained in a pre-defined set, i.e., the inequality −ρ1ðtÞ < e1ðtÞ < ρ1ðtÞ holds.
Proof. Define the tracking error of system (3) as
e1 = x1c − x1,
e2 = x2c − x2:ð32Þ
As the control command x1c is assumed as a constant, wehave x2c = _x1c = 0. The time derivative of e1 and e2 along (3) is
_e1 = _x1c − _x1 = −x2 = e2, ð33Þ
_e2 = _x2c − _x2 = − _x2 = −f − bu: ð34ÞBased on the expression of the transformed error ε1, it
yields
ε1 =12ln 1 + λ1ð Þ − ln 1 − λ1ð Þ½ �: ð35Þ
Then, the function ln ð1 + λ1Þ and ln ð1 + λ1Þ can bederived using Taylor’s expansion as follows:
ln 1 + λ1ð Þ = λ1 −12λ21 + R2 λ1ð Þ, ð36Þ
ln 1 − λ1ð Þ = −λ1 +12λ21 + R2 λ1ð Þ, ð37Þ
where R2ðλ1Þ is the 2-order Taylor remainder. As the error λ1is relatively small near zero, thus the R2ðλ1Þ can be neglected.
Substituting (36) into (35), the transformed error ε1 canbe approximately expressed as
ε1 = λ1 =x1c − z1ρ1 tð Þ : ð38Þ
Substituting (38) into (19), the proposed controller (18)can be rewritten as
u =kp x1c − z1/ρ1 tð Þð Þ − kdz2 − z3
b: ð39Þ
Substituting (39) into (34) yields
_e2 = −f − bu = −kpx1c − z1ρ tð Þ + kd z2 + z3 − f : ð40Þ
Combining the expressions of the observer error and sys-tem error, we can obtain that
From (33) and (44), the tracking error model is given by
_e =Ae e + Be ez , ð45Þ
where e = ½e1, e2�T and ez = ½ez1, ez2�T are the system trackingerror vector and the observer error vector, respectively. Thematrix Ae and Be are described as
Ae =0 1
−kp
ρ1 tð Þ −kd
264
375,
Be =0 0 0
−kp
ρ1 tð Þ −kd −1
264
375:
ð46Þ
6 International Journal of Aerospace Engineering
Meanwhile, the prescribed performance function ρ1ðtÞsatisfies
limt→∞
ρ1 tð Þ = limt→∞
ρ10 − ρ1∞ð Þ exp −κ1tð Þ + ρ1∞½ � = ρ1∞:
ð47Þ
Thus, under the condition t⟶∞, the matrix Ae and Becan be given as
Ae =0 1
−kpρ1∞
−kd
264
375,
Be =0 0 0
−kpρ1∞
−kd −1
264
375:
ð48Þ
According to Theorem 1, we can obtain that limt→∞
kezk = 0.Furthermore, if the characteristic polynomial s2 + kds + kp/ρ1∞ satisfies the Routh criterion, here, we select s2 + kds +kp/ρ1∞ = ðs + ωcÞ2, where ωc > 0 is the controller bandwidth,i.e., kp = ρ∞ω2
c , kd = 2ωc such that the matrix Ae is Hurwitz.Based on Lemma 1, we can obtain that the lim
t→∞eiðtÞ = 0, i =
1, 2 holds, i.e., the system (3) is asymptotically stable.Besides that, we can obtain the following equation by
conducting the inverse operation based on the transformederror (15) as
exp 2ε1ð Þ = 1 + λ11 − λ1
: ð49Þ
Based on Lemma 2, if the transformed errors ε1 = T−1ðλ1Þ can be controlled to be bounded, i.e., jε1j ≤ εM1 holdsfor positive constants εM1 > 0. This further implies
−1 <exp −εM1ð Þ − 1exp −εM1ð Þ + 1
≤ λ1 ≤exp εM1ð Þ − 1exp εM1ð Þ + 1
< 1: ð50Þ
From the fact λ1 = e1ðtÞ/ρ1ðtÞ and ρ1ðtÞ > 0, the error e1can be maintained within a predefined set, −ρ1ðtÞ < e1ðtÞ <ρ1ðtÞ holds. This completes the proof.
4. Simulation Results and Analysis
In order to evaluate the performance of the proposed ASVattitude controller, three numerical simulation cases are con-ducted in different configurations and scenarios.
4.1. Contrast Scheme. In order to demonstrate effectivenessand superiority of the proposed controller with the pre-scribed performance control-based LADRC (LADRC-PPC)method, the LADRC method [30] is introduced to the simu-
lation scenarios for comparison study, and the controller isdevised as follows:
uLADC =kp x1c − z1ð Þ − kdz2 − z3
b, ð51Þ
where z1, z2, and z3 is obtained by LESO.
4.2. Flight Conditions. The vehicle is assumed to maintainaltitude H = 12 km cruise at constant velocity V = 1000m/ssuch that the flight path angle is θ = 0 deg, i.e., the variationof the angle of attack α is the same as that of the pitch angleϑ based on the system (1). The aerodynamic coefficients arem0 = 0:0005, mα
z = −0:018 deg−1, mα2z = −0:003 deg−2 , and
mδzz = −0:035 deg−1. The initial conditions for the ASV are
set as ϑð0Þ = 0 deg, ωzð0Þ = 0 deg/s.
4.3. Controller Parameters. The relevant parameters of PPFis designed as ρ10 = 1:5, ρ1∞ = 0:05, and κ1 = 5, i.e., thePPF is selected as ρ1ðtÞ = ð1:5 − 0:05Þ exp ð−5tÞ + 0:05. Theobserver bandwidth ω0 is set as ω0 = 40, and controllerbandwidth ωc is designed as ωc = 10. Then, the observergains β1, β2, and β3 and controller gains kp, kd can beobtained according to the correspondence with the band-width ω0 and ωc.
4.4. Evaluation Index. For quantitatively contrasting the per-formance, the following performance indexes are introducedto evaluate the above control schemes.
(1) Convergence time (CT) tc of the tracking error (TE)
The convergence time tc is assumed as the corre-sponding time that the tracking error e1 is e1 ≤ 10−6and continues at least 10 sampling periods.
where t0 and t f are the start time and end time,respectively, u is the control input.
7International Journal of Aerospace Engineering
Case 1. This simulation is conducted in a standard condition.The aerodynamic coefficients are set to be the nominal valueswithout external disturbance, i.e., the term Δ in system (3) isΔ = 0.
The comparative results are summarized in Table 1 andFigures 3–10.
As depicted in Figures 3 and 4, the ASV can track thedesired attitude command with the listed control approaches
Table 1: The comparative simulation results of Case 1.
Index LADRC LADRC-PPC Ratio
CT tc 2.65 s 0.99 s 2.57 : 1
AV μe 0.07 deg 0.03 deg 2.33 : 1
SD σe 0.18 deg 0.14 deg 1.28 : 1
ACC Q 1.26 1.30 1 : 1.03
1.2
1
0.8
0.6
0.4
0.2
𝜗 (d
eg)
00 1 2 3
t (s)4
LADRCLADRC-PPC
5
Figure 3: Pitch angle.
Trac
king
erro
r of 𝜗
(deg
)
t (s)
1.5
1
0.5
0
–0.5
–1
–1.50 1 2 3 4 5
LADRCLADRC-PPCPPF
Figure 4: Tracking error of pitch angle.
𝜔z (
deg/
s)
t (s)
5
4
3
2
1
0
–10 1 2 3 4 5
LADRCLADRC-PPC
Figure 5: Pitch angular rate.
𝛿z (
deg)
t (s)
0
–0.2
–0.4
–0.6
–0.8
–1
–1.2
–1.40 1 2 3 4 5
LADRCLADRC-PPC
Figure 6: Control input signal-elevator deflection.
0
–0.2
–0.4
–0.6
–0.8
–10 1 2 3
t (s)4
Actual - LADRC
5
Estimated - LADRC
f (d
eg/s
2 )
Figure 7: Estimation of f with LADRC.
8 International Journal of Aerospace Engineering
accurately. It is worth noting that the tracking error is limitedin a prescribed set with a fairly satisfactory tracking errorresponse by utilizing the presented LADRC-PPC controller.As listed in Table 1, the tracking error remains remarkablysmall with a faster rate and a much smaller average errorwhich the average error is 0.03 deg, and converges to theneighbourhood of zero in approximately 0.99 s with theLADRC-PPC controller. However, the corresponding valueswith the LADRC scheme are 0.07 deg and 2.55 s, respectively.That is, the steady-state and transient performances of theASV attitude system under the designed LADRC-PPC con-troller is obviously superior than the contrast LADRC whichis owing to the help of introducing the PPC technology. Wecan see that initial value with the LADRC-PPC controller isslightly larger than the one with LADRC which is due tothe faster convergence in the initial phase with the LADRC-PPC approach as shown in Figures 5 and 6, and the amountof control consumption under the two schemes is virtuallyidentical, i.e., the novel proposed LADRC-PPC scheme canachieve an excellent control performance with approximatelythe same control consumption in comparison with theLADRC method. Figures 7 and 8 illustrate the estimationsof disturbances; it can be seen that LESO can precisely andrapidly estimate the total disturbances of the system, andthe convergence rate of LESO under the LADRC-PPC is fas-ter than that of LADRC as show in Figures 9 and 10.
Case 2. This case is performed in a perturbed condition toverify the robustness of the proposed method. The uncer-tainty term Δ is mainly considered two parts: (1) the modeluncertainties Δ1, and the aerodynamic coefficients are pre-sumed to decrease by 10% on the basis of standard values,the density of air ρ is perturbed to be +5% and (2) externaldisturbance Δ2, and it is considered an abrupt one Δ2 = 0:1deg/s2ðt ≥ 3Þ, i.e., the uncertainty term Δ is set as follows
Δ =Δ1 t < 3ð ÞΔ1 + 0:1 t ≥ 3ð Þ
( deg/s2: ð55Þ
The comparative results are summarized in Table 2 andFigures 11–18.
From Figure 11, the ASV attitude system can track thedesired command successfully by using the mentioned twocontrol approaches in the presence of unknown dynamics,uncertainties, and disturbances. The tracking error responseof the system with the proposed LADRC-PPC scheme is alsobetter than the classic LADRC method in Figure 12. Espe-cially, when the abrupt external disturbance occurs, thetracking error under the LADRC-PPC controller can con-verge to a compact set less than 0.5 s, and the correspondingconvergence time under the LADRC approach is nearly 2.0 s.That is, the transient performance under the LADRC-PPCcontroller is significantly improved than the contrast LADRCin spite of unknown uncertainties and disturbances. FromFigures 13 and 14, it can be seen that the variations of thestates under the perturbed condition are generally accordantwith the standard condition. Meanwhile, the amount of con-trol consumption under the two controller is similar. In
0.2
0
–0.2
–0.4
–0.6
–0.8
–10 1 2 3
t (s)4 5
Actual - LADRC-PPCEstimated - LADRC-PPC
f (d
eg/s
2 )
Figure 8: Estimation of f with LADRC-PPC.
0.05
0
–0.05
–0.1
–0.15
–0.20 1 2 3
t (s)4 5
LADRC
Estim
atio
n er
ror (
deg/
s2 )
Figure 9: Estimation error of f with LADRC.
0.05
0
–0.05
–0.1
–0.15
–0.2
–0.25
–0.30 1 2 3
t (s)4 5
LADRC-PPC
Estim
atio
n er
ror (
deg/
s2 )
Figure 10: Estimation error of f with LADRC-PPC.
9International Journal of Aerospace Engineering
addition, the two controllers can achieve an excellent track-ing performance owing to the total disturbances can be pre-cisely estimated with the LESO as shown in Figures 15–18.
Case 3. For further evaluating the robust performance of theproposed control scheme, a Monte Carlo analysis consisting
Table 2: The comparative simulation results of Case 2.
Index LADRC LADRC-PPC Ratio
CT tc 2.74 s 1.10s 2.49 : 1
AV μe 0.07 deg 0.03 deg 2.33 : 1
SD σe 0.18 deg 0.14 deg 1.28 : 1
ACC Q 1.06 1.10 1 : 1.04
1.2
1
0.8
0.6
0.4
0.2
𝜗 (d
eg)
00 1
1.03
1.02
1.01
1
3 3.5 4 4.5 5
2 3t (s)
4
LADRCLADRC-PPC
5
Figure 11: Pitch angle.
1.5
1
0.5
0
–0.5 –0.01
–0.02
–0.03
0
3 3.5 4.54 5
–1
–1.50 1 2 3
t (s)4
LADRCLADRC-PPCPPF
5
Trac
king
erro
r of 𝜗
(deg
)
Figure 12: Tracking error of pitch angle.
5
4
3
2
1
0
–10 1 2 3
t (s)4
LADRCLADRC-PPC
5
𝜔z (
deg/
s)
Figure 13: Pitch angular rate.
0
–0.5
–1
–1.5
0 1 2 3t (s)
4
LADRCLADRC-PPC
5
𝛿z (
deg)
Figure 14: Control input signal-elevator deflection.
0 1 2 3t (s)
4
Actual - LADRCEstimated - LADRC
5
0
–0.2
–0.4
–0.6
–0.8
f (d
eg/s
2 )
Figure 15: Estimation of f with LADRC.
10 International Journal of Aerospace Engineering
of N = 1000 sample runs is developed, and the parameterperturbations are considered
(1) There is a random variation between -20% and +20%obeying normal distribution for the aerodynamicscoefficients
(2) The perturbation of parameter Jz and ρ is -10%~+10% of the standard value, which follows the ran-dom variation with normal distribution
Figures 19–24 depicts the 1000-run Monte Carlo simula-tion results, and Table 3 illustrates the corresponding statisti-cal results including the expectation value μ and averagevariance σ of the above indexes CT tc, AV μe, and ACC Q.
The Monte Carlo simulation results are depicted inFigures 19–24, and Table 3 lists the statistical results of1000 Monte Carlo simulations which included the expecta-tion μ and average variance σ of the concerned indexes CTtc, AV μe, and ACC Q.
0
–0.2
–0.4
–0.6
–0.80 1 2 3
t (s)4 5
Actual - LADRC-PPCEstimated - LADRC-PPC
f (d
eg/s
2 )
Figure 16: Estimation of f with LADRC-PPC.
0.1
0
0.05
–0.05
–0.150 1 2 3 4 5
54
0
0.01
0.02
0.03
4.53.5
–0.1
t (s)
Estim
atio
n er
ror (
deg/
s2 )
LADRC
Figure 17: Estimation error of f with LADRC.
0.1
0
0.05
–0.05
–0.15
–0.1
–0.2
0.08
0.06
0.04
0.02
3.2 3.4 3.6 3.80
0 1 2 3 4 5t (s)
Estim
atio
n er
ror (
deg/
s2 )
LADRC-PPC
Figure 18: Estimation error of f with LADRC-PPC.
3.1
2.9
2.8
2.7t c (s
)
2.6
2.5
2.40 200 400 600
N (times)800 1000
3
LADRC
Figure 19: CT of Monte Carlo.
1.4
1.3
1.2
1.1
0.9
0.8
0.70 200 400 600 800 1000
1t c (s
)
N (times)
LADRC-PPC
Figure 20: CT of Monte Carlo.
11International Journal of Aerospace Engineering
Figures 19 and 20 depict the convergence time of thetracking error history for 1000 times with the mentionedtwo controllers, we can obtain that the expectation of CT tcwith the proposed LADRC-PPC 1.07 s that is shorter thanthe one of the classic LADRC approach, and it can achievea smaller dispersion with LADRC-PPC. The statistical resultsμe exhibit the similar regularity as shown in Figures 21 and22. The distributions of the amount of control consumptionQ with the above controllers are illustrated in Figures 23and 24, and the expectation of Q under LADRC-PPC is1.228, which is very close to the one of the LADRC method.The above results further verify that the proposed LADRC-PPC approach exhibits enhanced robustness despite multipledisturbances, and it can achieve an obviously superiorsteady-state and transient performances than the classicLADRC-PPC method.
5. Conclusion
In this paper, a novel LADRC scheme is proposed for thrASV attitude control system with multiple disturbances andprescribed performance constraint. The chief feature is thatit introduces the PPC technique into the LADRC design
𝜇e (
deg)
0.07
0.068
0.066
0.064
0.062
0.060 200 400 600
N (times)800 1000
LADRC
Figure 21: AV of Monte Carlo.
𝜇e (
deg)
0.041
0.04
0.039
0.038
0.037
0.036
0.0350 200 400 600 800 1000
N (times)
LADRC-PPC
Figure 22: AV of Monte Carlo.
Q
2.5
1.5
0.5
2
1
0 200 400 600N (times)
800 1000
LADRC
Figure 23: ACC of Monte Carlo.
Q
2.5
1.5
0.5
2
1
0 200 400 600 800 1000N (times)
LADRC-PPC
Figure 24: ACC of Monte Carlo.
Table 3: The statistical results of the Monte Carlo simulations.
Statistical data LADRC LADRC-PPC
tcμ 2.7257 1.0710
σ 0.1022 0.0750
μe
μ 0.0641 0.0374
σ 0.0015 0.0007
Q
μ 1.2270 1.2880
σ 0.2253 0.2353
12 International Journal of Aerospace Engineering
process so that the tracking error can be strictly confined toan adjustable residual set with the prescribed steady-stateand transient performances. Meanwhile, the LESO isemployed to estimate the total disturbances such that it caneffectively solve the unknown model dynamic problem. The-oretical analysis and contrast simulation results validate thata superior robust tracking performance of the ASV attitudesystem with the proposed control strategy design. Further-more, both the transient and steady-state performances ofthe classical LADRC can be significantly improved by intro-ducing the PPC method. Future work will consider thecompletely unknown dynamic control system design withactuator saturation via the presented control approach.
Nomenclature
ϑ: Pitch angleωz : Pitch angular rateθ: Flight path angleα: Angle of attackJz : Pitch moment of inertiaMz : Pitch momentS: Reference areaL: Reference length�q = 0:5ρV2: Dynamic pressureρ: Density of airV : Velocityδz : Elevator deflectionexp ðÞ: Exponential function with natural constant.
Data Availability
The data used to support the findings of this study are avail-able from the corresponding author upon request.
Conflicts of Interest
The authors declared no potential conflicts of interest withrespect to the research, authorship, and/or publication of thisarticle.
Acknowledgments
The authors are grateful for the projects supported by theNational Natural Science Foundation of China (Grant No.11176012).
References
[1] B. H. Dauby, D. W. Adamczak, J. A. Muse, and M. A. Bolen-der, “HIFiRE 6: overview and status update 2014,” in 20thAIAA International Space Planes and Hypersonic Systemsand Technologies Conference, pp. 1–17, Glasgow, Scotland,July 2015.
[2] X. W. Bu, Q. Wang, Y. Zhao, and G. J. He, “Concise neuralnonaffine control of air-breathing hypersonic vehicles subjectto parametric uncertainties,” International Journal of Aero-space Engineering, vol. 2017, pp. 1–16, 2017.
[3] A. Maity and R. Padhi, “Robust control design of an air-breathing engine for a supersonic vehicle using backstepping
and UKF,” Asian Journal of Control, vol. 19, no. 5, pp. 1710–1721, 2017.
[4] C. Y. Xu, H. M. Lei, and N. Na, “Active disturbance rejectioncontrol for air-breathing hypersonic vehicles based on pre-scribed performance function,” International Journal of Aero-space Engineering, vol. 2019, pp. 1–20, 2019.
[5] D. O. Sigthorsson, P. Jankovsky, A. Serrani, S. Yurkovich,M. A. Bolender, and D. B. Doman, “Robust linear output feed-back control of an airbreathing hypersonic vehicle,” Journal ofGuidance, Control, and Dynamics, vol. 31, no. 4, pp. 1052–1066, 2008.
[6] G. B. Cai, G. R. Duan, and C. H. Hu, “A velocity-based LPVmodeling and control framework for an air-breathing hyper-sonic vehicle,” International Journal of Innovative Comput-ing, Information and Control, vol. 7, no. 5, pp. 2269–2281,2011.
[7] Z. Cheng, P. Xue, and Y. L. Liu, “Nonlinear controller designfor hypersonic vehicles based on dynamic inversion consider-ing flexible states,” Advanced Materials Research, vol. 424-425,pp. 701–708, 2012.
[8] G. Wu, X. Meng, and F. Wang, “Improved nonlinear dynamicinversion control for a flexible air-breathing hypersonic vehi-cle,” Aerospace Science and Technology, vol. 78, pp. 734–743,2018.
[9] G. D. Zhu and Z. J. Shen, “Three dimensional trajectory linear-ization control for flight of air-breathing hypersonic vehicle,”Procedia Engineering, vol. 99, pp. 1108–1119, 2015.
[10] Z. Pu, X. Tan, G. Fan, and J. Yi, “Uncertainty analysis androbust trajectory linearization control of a flexible air-breathing hypersonic vehicle,” Acta Astronautica, vol. 101,pp. 16–32, 2014.
[11] X. Cheng, P. Wang, and G. Tang, “Fuzzy-reconstruction-basedrobust tracking control of an air-breathing hypersonic vehi-cle,” Aerospace Science and Technology, vol. 86, pp. 694–703,2019.
[12] L. Dou, P. Su, Q. Zong, and Z. Ding, “Fuzzy disturbanceobserver-based dynamic surface control for air-breathinghypersonic vehicle with variable geometry inlets,” IET ControlTheory and Applications, vol. 12, no. 1, pp. 10–19, 2018.
[13] A. B. Waseem, Y. Lin, and S. K. Amezquita, “Adaptivedynamic surface control of a hypersonic flight vehicle withmagnitude, rate and bandwidth constraints,” IFAC Proceed-ings, vol. 44, no. 1, pp. 5341–5346, 2011.
[14] X. Wang, J. Guo, S. Tang, and S. Qi, “Fixed-time disturbanceobserver based fixed-time back-stepping control for an air-breathing hypersonic vehicle,” ISA Transactions, vol. 88,pp. 233–245, 2019.
[15] X. Bu, G. He, and K. Wang, “Tracking control of air-breathinghypersonic vehicles with non-affine dynamics via improvedneural back-stepping design,” ISA Transactions, vol. 75,pp. 88–100, 2018.
[16] Y. Yang and Y. Yan, “Neural network approximation-basednonsingular terminal sliding mode control for trajectory track-ing of robotic airships,” Aerospace Science and Technology,vol. 54, pp. 192–197, 2016.
[17] L. Hu, R. Li, T. Xue, and Y. Liu, “Neuro-adaptive tracking con-trol of a hypersonic flight vehicle with uncertainties using rein-forcement synthesis,” Neurocomputing, vol. 285, pp. 141–153,2018.
[18] J. Wang, Q. Zong, R. Su, and B. Tian, “Continuous high ordersliding mode controller design for a flexible air- breathing
13International Journal of Aerospace Engineering
hypersonic vehicle,” ISA Transactions, vol. 53, no. 3, pp. 690–698, 2014.
[19] P. Yu, Y. Shtessel, and C. Edwards, “Continuous higher ordersliding mode control with adaptation of air breathing hyper-sonic missile,” International Journal of Adaptive Control andSignal Processing, vol. 30, no. 8-10, pp. 1099–1117, 2016.
[20] Z. Wang, W. Bao, and H. Li, “Second-Order dynamic sliding-mode control for nonminimum phase underactuated hyper-sonic vehicles,” IEEE Transactions on Industrial Electronics,vol. 64, no. 4, pp. 3105–3112, 2017.
[21] H. An, C. Wang, and B. Fidan, “Sliding mode disturbanceobserver-enhanced adaptive control for the air- breathinghypersonic flight vehicle,” Acta Astronautica, vol. 139,pp. 111–121, 2017.
[22] R. Parvathy and A. E. Daniel, “A survey on active disturbancerejection control,” in 2013 International Mutli-Conference onAutomation, Computing, Communication, Control and Com-pressed Sensing (iMac4s), pp. 330–335, Kottayam, India,March 2013.
[23] H. An, J. Liu, C. Wang, and L. Wu, “Disturbance observer-based antiwindup control for air-breathing hypersonic vehi-cles,” IEEE Transactions on Industrial Electronics, vol. 63,no. 5, pp. 3038–3049, 2016.
[24] Z. Q. Gao, “Scaling and bandwidth-parameterization basedcontroller tuning,” in Proceedings of the American ControlConference, pp. 4989–4996, Denver, CO, USA, USA, June2003.
[25] H. Niu, Q. Gao, Z. Liu, S. Tang, and W. Guan, “Improved lin-ear active disturbance rejection control for lever-type electricerection system with varying loads and low-resolutionencoder,” Mathematical Problems in Engineering, vol. 2019,13 pages, 2019.
[26] R. Morales, H. Sira-Ramírez, and J. A. Somolinos, “Linearactive disturbance rejection control of the hovercraft vesselmodel,” Ocean Engineering, vol. 96, pp. 100–108, 2015.
[27] L. Zhao, B. Zhang, H. Yang, and Q. Li, “Optimized linear activedisturbance rejection control for pneumatic servo systems vialeast squares support vector machine,” Neurocomputing,vol. 242, pp. 178–186, 2017.
[28] S. Ali, G. Yang, and C. Huang, “Performance optimization oflinear active disturbance rejection control approach by modi-fied bat inspired algorithm for single area load frequency con-trol concerning high wind power penetration,” ISATransactions, vol. 81, pp. 163–176, 2018.
[29] J. J. Carreño-Zagarra, J. L. Guzmán, J. C. Moreno, andR. Villamizar, “Linear active disturbance rejection control fora raceway photobioreactor,” Control Engineering Practice,vol. 85, pp. 271–279, 2019.
[30] H. Li, F. Yu, F. Li, S. Chuang, and Y. Zhide, “Missile attitudecontrol in a large flight envelope based on the linear active dis-turbance rejection control approach,” in Proceedings of 2014IEEE Chinese Guidance, Navigation and Control Conference,pp. 1219–1223, Yantai, China, August 2014.
[31] H. An and Q. Wu, “Disturbance rejection dynamic inversecontrol of air-breathing hypersonic vehicles,” Acta Astronau-tica, vol. 151, pp. 348–356, 2018.
[32] J. Na, Y. Huang, X. Wu, G. Gao, G. Herrmann, and J. Z. Jiang,“Active adaptive estimation and control for vehicle suspen-sions with prescribed performance,” IEEE Transactions onControl System Technology, vol. 26, no. 6, pp. 2063–2077,2018.
[33] Y. Huang, J. Na, X. Wu, and G. Gao, “Approximation-Freecontrol for vehicle active suspensions with hydraulic actuator,”IEEE Transactions on Industrial Electronics, vol. 65, no. 9,pp. 7258–7267, 2018.
[34] Z. Zhou, Y. Zhang, and D. Zhou, “Robust prescribed perfor-mance tracking control for free-floating space manipulatorswith kinematic and dynamic uncertainty,” Aerospace Scienceand Technology, vol. 71, pp. 568–579, 2017.
[35] J. Na, Q. Chen, X. Ren, and Y. Guo, “Adaptive prescribed per-formance motion control of servo mechanisms with frictioncompensation,” IEEE Transactions on Industrial Electronics,vol. 61, no. 1, pp. 486–494, 2014.
[36] S. Wang, J. Na, and X. Ren, “RISE-based asymptotic prescribedperformance tracking control of nonlinear servo mecha-nisms,” IEEE Transactions on Systems, Man and Cybernetics:Systems, vol. 48, no. 12, pp. 2359–2370, 2018.
[37] Y. Zhang, S. Wang, B. Chang, and W. H. Wu, “Adaptive con-strained backstepping controller with prescribed performancemethodology for carrier-based UAV,” Aerospace Science andTechnology, vol. 92, pp. 55–65, 2019.
[38] C. Ming, R. Sun, and B. Zhu, “Nonlinear fault-tolerant controlwith prescribed performance for air-breathing supersonic mis-siles,” Journal of Spacecraft and Rockets, vol. 54, no. 5,pp. 1092–1099, 2017.
[39] C. Y. Xu, H. M. Lei, and N. Na, “Active disturbance rejectioncontrol for air-breathing hypersonic vehicles based on pre-scribed performance function,” International Journal of Aero-space Engineering, vol. 2018, pp. 1–12, 2018.
[40] C. Zhang, G. Ma, Y. Sun, and C. Li, “Observer-based pre-scribed performance attitude control for flexible spacecraftwith actuator saturation,” ISA Transactions, vol. 89, pp. 84–95, 2019.
[41] X. Bu, X. Wu, F. Zhu, J. Huang, Z. Ma, and R. Zhang, “Novelprescribed performance neural control of a flexible air-breathing hypersonic vehicle with unknown initial errors,”ISA Transactions, vol. 59, pp. 149–159, 2015.
[42] X. Bu, Y. Xiao, and K. Wang, “A prescribed performance con-trol approach guaranteeing small overshoot for air-breathinghypersonic vehicles via neural approximation,” Aerospace Sci-ence and Technology, vol. 71, pp. 485–498, 2017.
[43] Y. Wang and J. Hu, “Improved prescribed performance con-trol for air-breathing hypersonic vehicles with unknown dead-zone input nonlinearity,” ISA Transactions, vol. 79, pp. 95–107, 2018.
[44] T. Berger, “Tracking with prescribed performance for linearnon-minimum phase systems,” Automatica, vol. 115, pp. 1–11, 2020.
[45] C. P. Bechlioulis and G. A. Rovithakis, “Robust approximationfree prescribed performance control,” in Proceedings of the19th Mediterranean Conference on Control and Automation,pp. 521–526, Corfu, Greece, June 2011.
[46] M. A. Bolender and D. B. Doman, “Nonlinear longitudinaldynamical model of an air-breathing hypersonic vehicle,”Journal of Spacecraft and Rockets, vol. 44, no. 2, pp. 374–387,2007.
[47] J. Parker, A. Serrani, S. Yurkovich, M. Bolender, andD. Doman, “Control-oriented modeling of an air-breathinghypersonic vehicle,” Journal of Guidance Control and Dynam-ics, vol. 30, no. 3, pp. 856–869, 2007.
[48] Q. Zheng, L. Q. Gao, and Z. Q. Gao, “On stability analysis ofactive disturbance rejection control for nonlinear time-
14 International Journal of Aerospace Engineering
varying plants with unknown dynamics,” in Proceedings of the46th IEEE Conference on Decision and Control, pp. 3501–3506,New Orleans, LA, USA, December 2007.
[49] S. Shao and Z. Gao, “On the conditions of exponential stabilityin active disturbance rejection control based on singular per-turbation analysis,” International Journal of Control, vol. 90,no. 10, pp. 2085–2097, 2016.
