Adaptive tracking control for a class of uncertain nonlinear systems with infinite number of actuator failures using neural networks

Advances in Difference Equations, Dec 2017

We consider adaptive compensation for infinite number of actuator failures in the tracking control of uncertain nonlinear systems. We construct an adaptive controller by combining the common Lyapunov function approach and the structural characteristic of neural networks. The proposed control strategy is feasible under the presupposition that the systems have a nonstrict-feedback structure. We prove that the states of the closed-loop system are bounded and the tracking error converges to a small neighborhood of the origin under the designed controllers, even though there are an infinite number of actuator failures. At last, the validity of the proposed control scheme is demonstrated by two examples.

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

https://advancesindifferenceequations.springeropen.com/track/pdf/10.1186/s13662-017-1426-5?site=advancesindifferenceequations.springeropen.com

Adaptive tracking control for a class of uncertain nonlinear systems with infinite number of actuator failures using neural networks

Lv and Wang Advances in Difference Equations Adaptive tracking control for a class of uncertain nonlinear systems with infinite number of actuator failures using neural networks Wenshun Lv Fang Wang We consider adaptive compensation for infinite number of actuator failures in the tracking control of uncertain nonlinear systems. We construct an adaptive controller by combining the common Lyapunov function approach and the structural characteristic of neural networks. The proposed control strategy is feasible under the presupposition that the systems have a nonstrict-feedback structure. We prove that the states of the closed-loop system are bounded and the tracking error converges to a small neighborhood of the origin under the designed controllers, even though there are an infinite number of actuator failures. At last, the validity of the proposed control scheme is demonstrated by two examples. nonlinear systems; actuator failures; adaptive control; backstepping; neural networks; nonstrict feedback 1 Introduction In recent years, many approximation-based adaptive fuzzy or neural backstepping controllers have been developed for uncertain nonlinear systems; see [–]. Among them, to eliminate the problem of ‘explosion of complexity’ inherent in the existing method, in [] a control design strategy was developed for a class of nonlinear systems in strict-feedback form with arbitrary uncertainty. To deal with the state unmeasured problem, a novel control scheme was introduced in []. To address the control problem of nonsmooth hysteresis nonlinearity in the actuator, adaptive neural controllers were constructed for nonlinear strict-feedback systems with unknown hysteresis in []. It should be noted that the control schemes mentioned are under the presupposition that the systems have a nonstrictfeedback structure. In the nonlinear systems without strict-feedback structure, the unknown nonlinear functions involve all the state variables, so they cannot be approximated with current states. To deal with such a structural restriction, in [] a variable separation method was proposed. The control scheme in [] assured that the tracking performance is achieved as time variable goes to infinity. Besides the proposed control scheme, many efforts have been made in relaxing such a restriction of system structure; see [–]. In practical application, the actuator component is usually employed to execute control actions on the plant. However, the actuation mechanism may suffer from failures, which results in the actuator losses of partial or total effectiveness. To prevent the emergence of performance deterioration and instability of the closed-loop system caused by actuator faults, accommodating actuator failures should be taken into account in the control design. In recent years, many control schemes have been proposed to accommodate actuator failures; see, for example, [–]. By applying backstepping technique for the linear systems, a systematic actuator failure compensation control was presented in []. Then, in [] the proposed control method was extended to nonlinear systems with actuator failures; in [] the problem of accommodating actuator failures was investigated for a lass of uncertain nonlinear systems with hysteresis input as a follow-up extension. In practice, the failure pattern in an actuator may change repeatedly, which makes failure parameters suffer from an infinite number of jumps. Consequently, the considered Lyapunov function would experience infinite number of jumps. In [], this problem was addressed by applying a new tuning function under the frame of adaptive control. However, the proposed control strategy can only apply to the strict-feedback systems. Motivated by the aforementioned researches, in this paper, we focus on the problem of adaptive compensation for an infinite number of actuator failures in neural tracking control for a class of nonstrict-feedback systems. The main contributions in this paper can be summarized as follows. () The control scheme in this paper relaxes the restriction of system structure so that a better approach is proposed to deal with the problem of compensation for an infinite number of actuator failures, which is more meaningful in practical application in comparison with []. () In this paper, combining neural networks and a new piecewise Lyapunov function analysis, we establish an adaptive control scheme for a class of uncertain nonlinear systems with a nonstrict-feedback structure. The remainder of the paper is organized as follows. In the next section, the problem description and preliminaries are presented. Section  shows the major result. In Section , the simulation result expounds the validity of the proposed control scheme. Finally, we give a simple summary. 2 Preliminaries and problem description 2.1 A. System description We consider the following nonstrict-feedback system form: x˙i = xi+ + fi(x),  ≤ i ≤ n – , m x˙n = bjσj(x)uj + fn(x), j= where x = [x, x, . . . , xn]T are the state vectors, fi (i = , , . . . , n) are unknown smooth nonlinear functions, uj ∈ R for j = , . . . , m is the output of the jth actuator, y denotes the system output, bj (j = , . . . , m) are unknown control coefficients, and σj(x) (j = , . . . , m) are known continuous functions. For simplicity, the internal dynamics in actuators can be negligible. Let uˇ j (j = , . . . , m) be jth actuator input, and then the normal actuator is described as uj = uˇ j. However, in practical applications, the actuation components may suffer from failures or faults. The actuator failure model in this paper is given as follows: uj = κj,k(t)uˇ j + ufj,k(t), κj,k(t)ufj,k(t) = , j = , . . . , m, () () where t ∈ [tk, tk+) for k = , , . . . , and tk denotes the unknown failure time instant at which the failure parameter jumps occur. In ()-(),  ≤ κj,k(t) ≤  is called the efficiency factor, and ufj,k(t) is the stuck value. Note that the proposed model covers two classes of typical actuator failures, that is, partial loss of effectiveness (PLOE) faults ( < κj,k < ) and total loss of effectiveness (TLOE) faults (κj,k = ), as detailed in []. Remark  The unknown time-varying value ufj,k(t) can be linearly parameterized as q ufj,k(t) = uj,k + i=jζji,kφji(t) with known functions φji(t) and unknown constants uj,k and ζji,k . Control objective: design the control input uˇ j to compensate for the actuator failures modeled as ()-() such that all the closed-loop signals are bounded and the system output y tracks the given reference signal yd with a tracking error converging to a residual around zero. The following assumptions are general in the literature on the adaptive actuator failure compensation control. Assumption  The number of failed actuators undergoing TLOE faults simultaneously is allowed to be at most m – , and achieving control objective with the remaining actuators is still possible. Assumption  The reference signal yd(t) and its first nth-order time derivatives yd(t) (i = , . . . , n) are known, smooth, and bounded. Assumption  sign(bj) are known for j = , , . . . , m. Assumption  The conditions σj(·) = ,  < b ≤ |bj| ≤ bM, and |ufj,k(t)| ≤ ufM are satisfied. In addition, for the PLOE faults,  < κ ≤ κj,k(t) < . Note that b, bM, ufM, and κ are known constants. Assumption  The failure parameters κj,k(t) and ufj,k(t) are smooth and continuous over [tk, tk+). Moreover, their change rates are bounded, that is, supt∈[tk,tk+) |κ˙j,k(t)| ≤ d and supt∈[tk,tk+) |u˙ fj,k(t)| ≤ d, where d and d denote two unknown positive constants. 2.2 B. RBF neural networks In the following design, the radial basis function (RBF) neural networks will be utilized to approximate an unknown function f (ζ ) defined on some compact set . For sufficiently large nodes number κ , the RBF neural networks ξ ∗T (ζ ) can approximate any continuous function f (ζ ) over a compact set ⊂ Rq to arbitrary accuracy ε >  as follows: f (ζ ) = ξ ∗T ϕ(ζ ) + (ζ ), ∀ζ ∈ ⊂ Rq, ξ ∗ := arg min sup f (ζ ) – ξ T ϕ(ζ ) , ξ∈Rκ ζ ∈ ϕi(ζ ) = exp – (ζ – ιi)ωTi(ζ – ιi) , i = , , . . . , κ, where ξ ∈ Rκ denotes the weight vector. In this research, the following Gaussian basis functions φi(ζ ) will be used: where ιi = [ιi, ιi, . . . , ιiq]T denotes the center of the receptive field, and ωi represents the width of the Gaussian function. Lemma  ([]) Let f (χ ) be a continuous function defined on a compact set . Then, for every ε > , there exists a neural network θ T (χ ) such that sup f (χ ) – θ T (χ ) ≤ ε. χ∈ ω  ≤ |ω| – ω tanh ε ≤ δε with δ = .. Lemma  ([]) For any ω ∈ R and ε > , the following holds: where the approximation error (ζ ) satisfies | (ζ )| ≤ ε, ϕ(ζ ) = [ϕ(ζ ), ϕ(ζ ), . . . , ϕκ (ζ )]T is the basis function vector, and ξ ∗ = [ξ, ξ, . . . , ξκ ]T ∈ Rκ is defined as () () () () () 3 Adaptive tracking control design and stability analysis In this section, based on a backstepping technique and neural networks, we design an adaptive actuator failure compensation control scheme. The control goal of this manuscript is to establish an adaptive controller such that the system output y follows a desired reference signal yd. 3.1 A. Adaptive tracking control design Before implementing the design, a set of uncertain constants is defined as Wi = Mi θi  = MiθiT θi, i = , . . . , n, where θi = [θi,, . . . , θi,Ni ] and Ni are the weight vector and the number of neurons in ith hidden layer, respectively. The operator ·  represents the Euclidean norm of a column vector. From the definition we know that Wi is an unknown constant. The adaptive parameter Wˆ i is utilized to estimate Wi, and the estimated error is W˜ i = Wi – Wˆ i. The computation in adaptation mechanism utilized in the literature can be reduced from in=Mi to n. The coordinate transformations are defined as follows: z = x – yd, zi = xi – αi– – y(di–), i = , . . . , n, where αi– is an intermediate controller. where h, ℘, and ℘ are positive design constants, and τ ∈ R(+m)(+m) is chosen to be symmetric and positive definite (for convenience, we can select τ = diag{r, r, . . . , r+m} with positive constants ri), Proj(·) denotes the projection operator, and τ is a known convex compact set given by Proof Consider the following Lyapunov function candidate: V¯ k =  n  i= zi + l W˜  + h ε˜ +  m  j= |bj|κj,k(t)τ˜T τ–τ˜, Denote uˇ j = sigσn(jbj) τˆT v, where τˆ = [τˆ, τˆ,, . . . , τˆ,m] is the estimate of τ ∈ R+m designed latter, and v = [αn + y(dn), σ, . . . , σm]. To ensure the boundedness of the jumping size of the Lyapunov function at failure instants, we design the adaptation laws for updating parameter estimators with projection operation. The adaptive laws are selected as n Wˆ˙ i = i= lai zi – ℘Wˆ i, i = , , . . . , n, ε˙ˆ = n i= hzi tanh zςi – ℘εˆ, τ˙ˆ = Proj(– τ vzn), τˆ() ∈ τ , τ = (γ, γ,, . . . , γ,m) mbM ≤ bκ and γ,j ≤ bbMuκfM , j = , . . . , m ; αi– is the virtual controller designed as αi = –λizi – ai ziWˆ i – εˆ tanh zςi ,  where λi, ai, and l are positive design constants. With the developed projection-based tuning function approach, we further propose new piecewise Lyapunov function analysis to establish the closed-loop system stability. 3.2 B. Stability analysis Theorem  Consider the closed-loop adaptive system consisting of nonlinear plant () with actuator failures ()-() and the proposed control scheme ()-(). Moreover, the nonstrictfeedback nonlinear system () satisfies Assumptions -. Then: () All the signals of the closed-loop system are bounded; () The tracking error converges to a small neighborhood of zero. where ε˜i is used to estimate εi, ε˜i = εi – εˆi refers to the estimated error, and τ˜i = τi – τˆi is the estimated error of τi. We define τ (t) ∈ R+m by τ(t) = τ,j(t) = –  m j= |bj|κj,k (t) , bjufj,k (t) m j= |bj|κj,k (t) , where j = , . . . , m. Note that all ,  ∈ τ satisfy  –  ≤  – bκ  mbM  + mbbmκ ufM = .   From the related properties of a projection operator in [] we can derive that τˆ (t) ∈ We get τ˜ (t) = τ (t) – τˆ (t) ≤ . The time derivative of V¯ is V˙¯ k = z(z + α + ϕ) + zi(zi+ + αi + ϕi – α˙ i–) m j= m j= + φ˙ + zn |bj|κj,k (t)τˆ T v + bjσjufj,k (t) + ϕn – α˙ n– – y(dn) , n– j= where It follows that φ = l W˜  + h ε˜ +  jm= |bj|κj,k (t)τ˜ T τ–τ˜ . φ˙ = – l W˜ Wˆ˙ – h ε˜ε˙ˆ – |bj|κj,k (t)τ˜ T τ–τˆ˙ + m j= |bj|κ˙j,k (t) τ T τ–τ˜ +  ˜ m j= |bj|κj,k (t)τ˜ T τ–τ˙ . Define unknown functions as Pi x, Wˆ , εˆ, yd, . . . , y(di–) = ∂ αi– i– ∂ Wˆ l l j= aj zj – zi ai j= i zj ∂∂αWjˆ– – ∂ αi– ℘Wˆ ∂ Wˆ + ∂∂αεiˆ– i– j= hzj tanh zςj + h tanh zςi j= i zj ∂∂αεjˆ– – ∂∂αεiˆ– ℘εˆ, () () τ . () () () () n i= m j= It follows that n i= n i= – m Since fi(x) contain the unknown functions ϕi, they cannot be implemented in practice. According to Lemma , for any given constant εi > , there exists a neural network θiT i such that λizi + i=n ai ziW˜ + i=n  ai + n i= ziε˜ tanh zςi + n i= .ς ε f(x) = ϕ(x), fi x, Wˆ , εˆ, yd, . . . , y(di–) = zi– + ϕi – i– ∂αi– (xj+ + ϕj) ∂xj From ()-() and ()-() we have zi Pi – ∂∂αWiˆ– W˙ˆ – ∂∂αεiˆ– ε˙ˆ ≤ . Moreover, according to () and (), we have () () () () () () () () () () () |bj|κj,k(t)τ T v + bjσjufj,k(t) – αn – y(dn) = . m j= m j= V˙¯ k(t) ≤ (αi + fi) – |bj|κj,k(t)τ˜T vzn + φ˙ . |zi|fi(x) ≤ ai zi θi  iT i +  ai + |zi|ε.  Combining with the condition iT i ≤ Mi and Lemma , we have |zi|fi(x) ≤ ai ziW +  ai + ziε tanh zςi + .ς ε,  where W = max θ  T , . . . , θi  iT i , ε = max{ε, . . . , εn}. Substituting () and () into (), we have |bj|κj,k(t)τ˜T vzn + φ˙ . By Assumption  and by () and () we get τ ˙ ≤ τ˙,j ≤ mbMd bκ , mbMd + mbMufMd bκ , j = , . . . , m. where where Then, it follows that V˙¯ k(t) ≤ –AV¯ k + B, m |bj|κj,k(t) τ T τ–τ˜, j= λmax( τ–) ˜  = mbMd   – τ F + mbM – τ F . V (t) = V¯ (t)k, t ∈ [tk, tk+), k = , , . . . , To establish the closed-loop system stability for all the time under the case of actuator failures or faults, we need to consider the overall Lyapunov function defined as where t =  is the initial time instant. Note that V (t) is not a continuous function, and it experiences a sudden jump at each failure instant tk+ (k = , , . . .). The jumping size at instant tk is computed as + – V tk+ – V tk+ = jm= |bj|κj,k+(tk++) τ˜T tk++ τ–τ˜ tk++ ≤ m j= – mbM  |bj|κj,k+(tk–+) τ˜ T tk–+  τ–τ˜ tk+ – τ– F  = . Let H(t) = eat V (t). Then it follows that H˙ = Aeat V (t) + eat V˙ (t) ≤ BeAt . Integrating both sides of () over [tk , tk+), we have – + H tk+ ≤ H tk+ + BeAt dt. From ()-() we have H tk++ ≤ H tk+ + BeAt dt + eatk+ . tk+ tk tk+ tk  T T tℵ(,T) We denote by ℵ(t, T ) the number of jumps of the overall Lyapunov V (t) during (t, T ) for t ≥ . Let T = min{tk+ – tk }, k = , , . . . . Then it follows that H T – = H tℵ+(,T) + BeAt dt ℵ(tk , T )T ≤ T – tk , k = , . . . , ℵ(, T ) tk – T ≤ –ℵ(tk , T )T . ≤ V () + BeAt dt + eAT ea(tk –T) , ℵ(,T) k= Then we have eAT ℵ(,T) k= It follows that ea(tk –T) ≤ eAT e–aℵ(tk ,T)T ℵ(,T) k= =  – e–aT ℵ(,T)  – e–aT eaT . V T – ≤ V () – AB –  – e–aT e–aT + AB +  – e–aT , ∀T > . Then all closed-loop signals are bounded. Note that bound of the tracking error can be derived as lim t→∞ z(t) ≤ AB +   – e–AT . n i= ≤ V (t), and let T → ∞. The () () () () () () () () () Remark  Inequality () makes a vital contribution to the backstepping design because it builds the relation between xi and zi, which makes a backstepping-based design procedure viable. Remark  The adaptive parameters Wˆ i, εˆ, τˆ are utilized to estimate Wi, ε, τ , respectively. W˜ i = Wi – Wˆ i, ε˜ = – εˆ, and τ˜ = τ – τˆ denote the estimated errors. Note that the failure parameter κj,k is allowed to be time varying during each time interval [tk, tk+) for k = , , . . . and bj (j = , . . . , m) are unknown control coefficients. The instability cannot be ensured when [tk, tk+) and bj are not contained in the Lyapunov function. Remark  The failure-related parameters τ contained in the Lyapunov function () will undergo a sudden jump at unknown time instant tk , and it follows that decreasing of the Lyapunov function, shown as in (), is only valid at the time interval [tk, tk+) during which the Lyapunov function V¯ k(t) is differentiable. To establish the closed-loop system stability under the case of actuator failures or faults, we consider the overall Lyapunov function defined in (). x˙ = x + x,  – e–x x˙ = xx +   + e–x + bu + bu, Z(ξ , x, x) = –ξ + x + tan– x +  ++ xx xx +  + e–x  – e–x , where y is the system output, and uj (j = , ) are control inputs to the plant. The parameters b =  and b = . are unknown for controller design. The last subsystem system ξ˙ = Z(ξ , x, x) represents the zero dynamics for the system. From [, ] we have that ξ = Z(ξ , x, x) is input-to-state stable with respect to [x, x], which implies that the state variable ξ is bounded if x(t), x(t) ∈ ∞. The system is initialized as x() = ., x() = , ξ () = , and u() = u() = . In the simulation, the failure case studied is given by u = cos . t – hT uc u = uc, t ∈ hT , (h + )T , h = , , , . . . , u =  sin(.t), u = .uc, t ∈ hT , (h + )T , h = , , , . . . , where uˇj is the jth actuator input for j = , , and T =  s denotes the minimum length of time intervals between two successive jumps of failure parameters. It is seen that during the time interval [hT , (h + )T ), h = , , , . . . , the first actuator suffers from PLOE faults with a time-varying efficiency indicator κ,h(t) = cos(.(t – hT )), whereas the second actuator works normally, and during [hT , (h + )T ), h = , , , . . . , the first actuator output is stuck at an unknown time-varying value, that is, u =  sin(.t), whereas the second actuator loses % of the effectiveness of its input. The boundary information b = ., bM = , κ = ., and ufM =  is employed to construct the projection adaptation law (), and then the compact set τ in () is computed as τ = [τ, τ,, τ,]T |. ≤ τ ≤ ., –. ≤ τ, ≤ ., j = , ]. () To make all the signals bounded and y follow a given signal yd = . sin(.t), we select the scheme uˇ j = τˆ(α + y¨d) + τˆ, + τˆ,, j = , , with τ = I, where I denotes the  ×  identity matrix. The initial parameter estimates are set as Wˆ () = , εˆ() = ., and τˆ() = [., ., .]T . Figures - demonstrate the corresponding simulation results. 4.2 B. Example 2 We consider the subsystem of the cascade chemical reactor system as in [, ]: x˙ = x + φ + x , x˙ = φ + bu + bu, y = x. In the system, and φ = (q + VqR)Td – x +VTd (q + qR) – CρAdcαρλ e– R(x+Td) – Ea UA ρcρ V x + Td – Tjd , () UA φ = ρcρ V where q and q = q + qR are the flows of reactants, b = ρUcρAVqjVj and b = ρUcρAVqjVj denote the control coefficients, V represents the volume of the reactor. The system parameters are given in Table . The system is initialized as x() = ., x() = , x() = , and u() = u() = . In the simulation, the failure case studied is given by u = uc, u = uc, t ∈ hT , (h + )T , h = , , , . . . , u = .uc, u = , t ∈ hT , (h + )T , h = , , , . . . , () () where uˇ j is the jth actuator input for j = , , and T =  s denotes the minimum length of time intervals between two successive jumps of failure parameters; b = ., bM = , κ = ., and ufM =  are employed to construct the projection adaptation law (), and then the compact set τ in () is computed as τ = [τ, τ,, τ,]T |. ≤ τ ≤ ., –. ≤ τ, ≤ ., j = , ]. () To make all the signals bounded and y follow a given signal yd = , we select the scheme uˇ j = τˆ(α + y¨d) + τˆ, + τˆ,, j = , , with τ = I, where I denotes the  ×  identity matrix. The initial parameter estimates are set as Wˆ () = , εˆ() = ., and τˆ() = [., ., .]T . Figures - demonstrate the corresponding simulation results. 4.3 C. Conclusions In this papedr, we investigate the issue of adaptive finite-time tracking for a class of nonlinearity systems with hysteresis. On the basis of the approximation capability of fuzzy logic systems, we give an adaptive law and an intermediate control function. Moreover, we proved that the aforementioned approach can make the system SGPFS. In comparison with the tuning function control schemes in [–], which also focus on adaptive actuator failure compensation problem, the framework of our control is further simplified by using RBFNN, as seen in ()-(). Moreover, with respect to the previous fault-tolerant controllers such as [, ], our scheme additionally contains an optimized neural network adaptation mechanism, which renders that there are only two estimates, that is, Wˆ i and ε need to be computed online. The control scheme in [] is under the presupposition that the systems have a nonstrict-feedback structure. Our control scheme relaxes the restriction of system structure so that the method in this note is more meaningful. In this sense, the presented study is more computationally attractive and thus feasible in practical implementation in comparison with other existing results. Acknowledgements The authors greatly appreciate the reviewers suggestions, the editors’ encouragement, and Yan Li’s contribution to this paper. Yan Li is currently professor at Shandong University of Science and Technology. She made a significant contribution to this paper, including control scheme design, paper writing, and revision. This work is supported partially by the National Natural Science Foundation of China (Grant No. 61503223), in part by the Project of Shandong Province Higher Educational Science and Technology Program (J15LI09), in part by China Postdoctoral Science Foundation-funded project 2016M592140, in part by Shandong innovation postdoctoral program 201603066, and in part by the SDUST Research Fund (2014TDJH102). Competing interests The authors declare that they have no competing interests. Authors’ contributions All authors contributed equally to writing this paper. All authors read and approved the final manuscript. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. 1. Wang , F , Chen, B , Lin , C , Li, XH : Distributed adaptive neural control for stochastic nonlinear multiagent systems . IEEE Trans. Cybern . 47 ( 7 ), 1795 - 1803 ( 2017 ) 2. Wang , F , Liu, Z , Zhang, Y , Philip Chen , CL : Adaptive fuzzy control for a class of stochastic pure-feedback nonlinear systems with unknown hysteresis . IEEE Trans. Fuzzy Syst . 24 ( 1 ), 140 - 152 ( 2016 ) 3. Tang , L , Zhao , J: Neural network based adaptive prescribed performance control for a class of switched nonlinear systems . Neurocomputing 230 , 316 - 321 ( 2017 ) 4. Shi , XC , Lu, JW , Li , Z , Xu, SY : Robust adaptive distributed dynamic surface consensus tracking control for nonlinear multi-agent systems with dynamic uncertainties . J. Franklin Inst . 353 , 4785 - 4802 ( 2016 ) 5. Wang , HQ , Chen, B , Liu, XP , Lin, C: Robust adaptive fuzzy tracking control for pure-feedback stochastic nonlinear systems with input constraints . IEEE Trans. Cybern . 43 ( 6 ), 2093 - 2104 ( 2013 ) 6. Kayacan , E , Ramonf, H , Saeys, W: Adaptive neuro-fuzzy control of a spherical rolling robot using sliding-mode-control-theory-based online learning algorithm . IEEE Trans. Cybern . 43 ( 1 ), 170 - 179 ( 2013 ) 7. Yu , JP , Ma, YM , Yu, HS , Lin, C: Reduced-order observer-based adaptive fuzzy tracking control for chaotic permanent magnet synchronous motors . Neurocomputing 214 , 201 - 209 ( 2016 ) 8. Hamayun , MT , Edwards, C , Alwi, H: Design and analysis of an integral sliding mode fault-tolerant control scheme . IEEE Trans. Autom. Control 57 ( 7 ), 1783 - 1789 ( 2012 ) 9. Tong, SC , Li, YM , Shi, P: Observer-based adaptive fuzzy backstepping output feedback control of uncertain MIMO pure-feedback nonlinear systems . IEEE Trans. Fuzzy Syst . 20 ( 4 ), 771 - 785 ( 2012 ) 10. Li , H , Chen , Z , Wu, L , Lam, HK , Du, H: Event-triggered fault detection of nonlinear networked systems . IEEE Trans. Cybern . 47 ( 4 ), 1041 - 1052 ( 2017 ) 11. Zhou , Q , Shi, P , Tian, Y , Wang, M: Approximation-based adaptive tracking control for MIMO nonlinear systems with input saturation . IEEE Trans. Cybern . 45 ( 10 ), 2119 - 2128 ( 2015 ) 12. Chen , B , Liu, XP , Liu, KF , Lin, C: Direct adaptive fuzzy control of nonlinear strict-feedback systems . Automatica 45 ( 6 ), 1530 - 1535 ( 2009 ) 13. Wang , D , Huang, J: Neural network-based adaptive dynamic surface control for a class of uncertain nonlinear systems in strict-feedback form . IEEE Trans. Neural Netw . 16 ( 1 ), 195 - 202 ( 2005 ) 14. Swaroop , D , Hedrick, J , Yip, P , Gerdes, J : Dynamic surface control for a class of nonlinear systems . IEEE Trans. Autom. Control 45 ( 10 ), 1893 - 1899 ( 2000 ) 15. Ge , SS , Wang , C : Adaptive neural control of uncertain MIMO nonlinear systems . IEEE Trans. Neural Netw . 15 ( 3 ), 674 - 692 ( 2004 ) 16. Chen , B , Liu, XP , Liu, K , Lin, C: Fuzzy-approximation-based adaptive control of strict-feedback nonlinear systems with time delays . Automatica 18 ( 5 ), 883 - 892 ( 2010 ) 17. Daniel , W , Li, JM , Niu, YG : Adaptive neural control for a class of nonlinearly parametric time-delay systems . IEEE Trans. Neural Netw . 16 ( 3 ), 625 - 635 ( 2005 ) 18. Zhang, TP, Ge, SS : Adaptive neural control of MIMO nonlinear state time-varying delay systems with unknown dead-zones and gain signs . Automatica 43 ( 6 ), 1021 - 1033 ( 2007 ) 19. Wang , F , Liu, Z , Yu, GY , Zhang, Y , Chen, X , Chen, CL : Adaptive neural control for a class of nonlinear time-varying delay systems with unknown hysteresis . IEEE Trans. Neural Netw. Learn. Syst . 25 ( 12 ), 2129 - 2140 ( 2014 ) 20. Chen , B , Liu, XP , Ge, SS , Lin , C: Adaptive fuzzy control of a class of nonlinear systems by fuzzy approximation approach . IEEE Trans. Fuzzy Syst . 20 ( 6 ), 1012 - 1021 ( 2012 ) 21. Liu, XP , Gu, GX , Zhou, KM : Robust stabilization of MIMO nonlinear systems by backstepping . Automatica 35 ( 2 ), 987 - 992 ( 1999 ) 22. Liu, XP , Jutan, A , Rohani , S: Almost disturbance decoupling of MIMO nonlinear systems and application to chemical processes . Automatica 40 ( 3 ), 465 - 471 ( 2004 ) 23. Ploycarpou , MM , Ioannou, PA: A robust adaptive nonlinear control design . Automatica 32 ( 3 ), 423 - 427 ( 1996 ) 24. Yao , B , Tomizuka , M: Adaptive robust control of SISO nonlinear systems in a semi-strict feedback form . Automatica 33 ( 5 ), 893 - 900 ( 1997 ) 25. Yao , B , Tomizuka , M: Adaptive robust control of MIMO nonlinear systems in semi-strict feedback forms . Automatica 37 ( 9 ), 1305 - 1321 ( 2001 ) 26. Tao , G , Joshi, SM , Ma, XL: Adaptive state feedback control and tracking control of systems with actuator failures . IEEE Trans. Autom. Control 46 ( 1 ), 78 - 95 ( 2001 ) 27. Tang , XD , Tao, G , Joshi, SM : Adaptive actuator failure compensation for parametric strict feedback systems and an aircraft application . Automatica 39 ( 11 ), 1975 - 1982 ( 2003 ) 28. Cai , JP , Wen, C , Su, HY , Li, XD : Robust adaptive failure compensation of hysteretic actuators for a class of uncertain nonlinear systems . IEEE Trans. Autom. Control 58 ( 9 ), 2388 - 2394 ( 2013 ) 29. Lai , GY , Liu, Z , Philip Chen , CL , Zhang, Y , Chen, X: Adaptive compensation for infinite number of time-varying actuator failures in fuzzy tracking control of uncertain nonlinear systems . IEEE Trans. Fuzzy Syst . ( 2017 ). doi: 10 .1109/TFUZZ. 2017 .2686338 30. Jajarmi , A , Hajipour, M , Baleanu , D: A new approach for the optimal control of time-varying delay systems with external persistent matched disturbances . J. Vib. Control. doi:10.1177/1077546317727821 31. Mobayen , S , Baleanu, D: Linear matrix inequalities design approach for robust stabilization of uncertain nonlinear systems with perturbation based on optimally-tuned global sliding mode control . J. Vib. Control 23 ( 8 ), 1285 - 1295 ( 2017 ) 32. Jajarmi , A , Hajipour, M , Baleanu , D: New aspects of the adaptive synchronization and hyperchaos suppression of a financial model . Chaos Solitons Fractals 99 , 285 - 296 ( 2017 ) 33. Wang , W , Wen, C: Adaptive actuator failure compensation control of uncertain nonlinear systems with guaranteed transient performance . Automatica 46 ( 12 ), 2082 - 2091 ( 2010 ) 34. Park , J , Sandberg, IW : Universal approximation using radial-basis-function network . Neural Comput . 3 ( 2 ), 246 - 257 ( 1991 ) 35. Ploycarpou , MM , Ioannou, PA: A robust adaptive nonlinear control design . Automatica 32 ( 3 ), 423 - 427 ( 1996 ) 36. Li , XJ , Yang, GH : Robust adaptive fault-tolerant control for uncertain linear systems with actuator failures . IET Control Theory Appl . 6 ( 10 ), 1544 - 1551 ( 2012 ) 37. Liu , X , Jutan , A , Rohani , S: Almost disturbance decoupling of MIMO nonlinear systems and application to chemical processes . Automatica 40 ( 3 ), 465 - 471 ( 2004 ) 38. Li , D: Neural network control for a class of continuous stirred tank reactor process with dead-zone input . Neurocomputing 131 , 453 - 459 ( 2014 ) 39. Tang , XD , Tao, G , Joshi, SM : Adaptive actuator failure compensation for parametric strict feedback systems and an aircraft application . Automatica 39 ( 11 ), 1975 - 1982 ( 2003 ) 40. Tang , XD , Tao, G , Joshi, SM : Adaptive actuator failure compensation for nonlinear MIMO systems with an aircraft control application . Automatica 43 ( 11 ), 1869 - 1883 ( 2007 ) 41. Cai , JP , Wen, C , Su, HY , Li, XD : Robust adaptive failure compensation of hysteretic actuators for a class of uncertain nonlinear systems . IEEE Trans. Autom. Control 58 ( 9 ), 2388 - 2394 ( 2013 ) 42. Wang , W , Wen, C: Adaptive actuator failure compensation control of uncertain nonlinear systems with guaranteed transient performance . Automatica 46 ( 12 ), 2082 - 2091 ( 2010 ) 43. Wang , CL , Wen, C , Lin, Y: Adaptive actuator failure compensation for a class of nonlinear systems with unknown control direction . IEEE Trans. Autom. Control 62 ( 1 ), 385 - 392 ( 2016 ) 44. Wang , CL , Wen, C , Lin, Y: Decentralized adaptive backstepping control for a class of interconnected nonlinear systems with unknown actuator failures . J. Franklin Inst . 352 ( 3 ), 835 - 850 ( 2015 ) 45. Tong , SC , Sui, S , Li, YM : Adaptive fuzzy decentralized tracking fault-tolerant control for stochastic nonlinear large-scale systems with unmodeled dynamics . Inf. Sci . 289 , 225 - 240 ( 2014 ) 46. Tong , SC , Wang, T , Li, YM : Fuzzy adaptive actuator failure compensation control of uncertain stochastic nonlinear systems with unmodeled dynamics . IEEE Trans. Fuzzy Syst . 22 ( 3 ), 563 - 574 ( 2014 ) 47. Miller , RH , William, BR : The effects of icing on the longitudinal dynamics of an icing research aircraft . In: 37th Aerospace Sciences, no. 99-0637 . AIAA, Washington ( 1999 )


This is a preview of a remote PDF: https://advancesindifferenceequations.springeropen.com/track/pdf/10.1186/s13662-017-1426-5?site=advancesindifferenceequations.springeropen.com

Wenshun Lv, Fang Wang. Adaptive tracking control for a class of uncertain nonlinear systems with infinite number of actuator failures using neural networks, Advances in Difference Equations, 374,