Switched-observer-based adaptive neural networks tracking control for switched nonlinear time-delay systems with actuator saturation

• Published in: Information sciences Vol. 621; pp. 36 - 57
• Main Authors: Li, Zhenhua, Cao, Gang, Xie, Wei, Gao, Rui, Zhang, Weidong
• Format: Journal Article
• Language: English
• Published: Elsevier Inc 01.04.2023
• Subjects: Actuator saturation; Adaptive neural networks (NNs); Average dwell time (ADT); Multiple Lyapunov–Krasovskii functionals (MLKFs); Output-feedback control; Switched nonlinear time-delay systems; Average dwell time (ADT); Multiple Lyapunov–Krasovskii functionals (MLKFs); Output-feedback control; Adaptive neural networks (NNs); Switched nonlinear time-delay systems; Actuator saturation
• ISSN: 0020-0255; 1872-6291
• DOI: 10.1016/j.ins.2022.11.094

This paper deals with adaptive neural networks (NNs) output-feedback tracking control problem for a general class of switched nonlinear time-delay systems with actuator saturation. To estimate unmeasurable states, a constructive delay-independent switched-observer-based adaptive NNs control approach for solving the problem is provided by using the average dwell time (ADT) method and the backstepping technique. Design obstacles stemmed from unknown time delays and actuator saturation and unknown nonlinear functions are overcome by using appropriate multiple Lyapunov–Krasovskii functionals method, the Nussbaum gain technique, and neural network approximations, respectively. The proposed control strategy guarantees that i) all the closed-loop signals for the switched system are bounded in the semi-global sense under a class of switching signals with ADT, attaining semi-global uniformly ultimately boundedness (SGUUB); and, ii) the output tracking error is driven towards, and kept within, a small neighbourhood around the origin. Simulation results verify the effectiveness of the proposed control strategy on two numerical examples, including a practical switched continuous stirred tank reactor (CSTR) system.