Sliding-mode surface-based adaptive actor-critic optimal control for switched nonlinear systems with average dwell time

Zhang, Haoyan; Wang, Huanqing; Niu, Ben; Zhang, Liang; Ahmad, Adil M.

doi:10.1016/j.ins.2021.08.062

Cited by 143 publications

(58 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Remark In most existing literature, under the so‐called universal approximation theorem, fuzzy logic systems 60 and neural networks 61,62 are commonly utilized to approximate virtual control functions and its differential items. Although the applicability of this method is strong, the control performance is still less than satisfactory.…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Event‐triggered adaptive tracking control for uncertain fractional‐order nonstrict‐feedback nonlinear systems via command filtering

Wang

Zhao

et al. 2022

Intl J Robust & Nonlinear

Self Cite

View full text Add to dashboard Cite

This article puts forward an event-triggered adaptive neural tracking control strategy for a class of uncertain fractional-order nonstrict-feedback nonlinear systems with full-state constraints and dead-zone inputs. The proposed control strategy is capable of surmounting an inherent shortcoming in most traditional backstepping control approaches which results from the repeated differentiation of virtual control functions, and enhancing the transmission efficiency of control signals by introducing an event-triggered mechanism with a dynamic threshold condition. Furthermore, by constructing a novel barrier Lyapunov function in each step of the backstepping procedure, the design difficulty arising from the system state constraints issue under fractional-order cases is resolved. It is strictly proved that the tracking error can converge into a desired small neighborhood of the origin and all signals in the closed-loop system are bounded. Finally, a simulation example is given to verify the effectiveness of our proposed control strategy.

show abstract

Section: Resultsmentioning

confidence: 99%

“…Remark 4. In most existing literature, under the so-called universal approximation theorem, fuzzy logic systems 60 and neural networks 61,62 are commonly utilized to approximate virtual control functions and its items.…”

Section: Stability Analysismentioning

confidence: 99%

Event‐triggered adaptive tracking control for uncertain fractional‐order nonstrict‐feedback nonlinear systems via command filtering

Wang

Zhao

et al. 2022

Intl J Robust & Nonlinear

Self Cite

View full text Add to dashboard Cite

show abstract

“…During the past decades, switched systems have received much attentions since they are widely applied in engineering practices (see References 1‐6 and the references therein). As a special kind of hybrid systems, switched system consists of a group of subsystems and a switching signal which specifies an active subsystem at each time instant.…”

Section: Introductionmentioning

confidence: 99%

Command filtering‐based adaptive neural network control for uncertain switched nonlinear systems using event‐triggered communication

Chen

Niu

Zhang

et al. 2022

Intl J Robust & Nonlinear

Self Cite

View full text Add to dashboard Cite

In this article, a command filtering‐based adaptive event‐triggered neural network control scheme is proposed for a class of uncertain switched nonlinear systems with unknown control coefficient and input saturation. First, radial basis function neural networks are used as function approximators to estimate unknown nonlinear functions. Then, an event‐triggering mechanism based on the tracking error is introduced to avoid the over‐consumption of communication resources. Furthermore, command filters are employed to solve the problem of complexity explosion that exists in conventional backstepping control design, and the error compensation signals are designed to reduce the errors caused by the filters. Considering that the unknown control gain and input saturation exist in many practical applications, a Nussbaum‐type function is thus introduced into the controller design to address these challenging issues. Finally, stability of the closed‐loop system is strictly proven under a standard Lyapunov stability analysis framework. The effectiveness of the proposed control scheme is illustrated by a simulation example.

show abstract

“…Thanks to the developments of neural networks recently, an approximation method called as adaptive dynamic programming (ADP) has attracted great attention by combining the DP with neural networks. [9][10][11][12][13][14][15] Through the ADP, neural networks are employed to approximate unknown functions and the policy iteration (PI) or value iteration (VI) schemes are used during the iteration process to obtain optimal solutions. In Reference 4, the PI-based ADP method was explored and the convergence and stability properties were also given.…”

Section: Introductionmentioning

confidence: 99%

Quantized optimal output feedback control and optimal triggering signal co‐design for unknown discrete‐time nonlinear systems

Han

Hao

et al. 2022

Optim Control Appl Methods

View full text Add to dashboard Cite

A problem of optimizing the cost function incorporated into the transmission cost is proposed for completely unknown discrete-time systems in this article. At first, a switched system containing two modes is formulated based on whether the triggering is executed or not, and the corresponding cost function and output feedback control law are also derived. By giving the formulated switched system, the triggering signal turns into the switching signal and the proposed problem is converted into an optimal control problem for the formulated discrete-time switched system. Secondly, a model network is presented to approximate the system dynamic since the dynamic model is completely unknown. Thirdly, in view of ADP-based optimal control methods of switched systems, critic and actor networks are applied respectively to approximate the optimal cost function and the optimal output feedback control law, and a PI-based online learning method is proposed. When weights of critic and actor networks are obtained and implemented through the proposed online learning method, the optimal triggering signal is derived. Fourthly, the convergent property is given to guarantee the obtained results are optimal. At last, the validity of the proposed method is shown by simulation results of two examples.

show abstract

Sliding-mode surface-based adaptive actor-critic optimal control for switched nonlinear systems with average dwell time

Cited by 143 publications

References 43 publications

Event‐triggered adaptive tracking control for uncertain fractional‐order nonstrict‐feedback nonlinear systems via command filtering

Event‐triggered adaptive tracking control for uncertain fractional‐order nonstrict‐feedback nonlinear systems via command filtering

Command filtering‐based adaptive neural network control for uncertain switched nonlinear systems using event‐triggered communication

Quantized optimal output feedback control and optimal triggering signal co‐design for unknown discrete‐time nonlinear systems

Contact Info

Product

Resources

About