State constraints‐based adaptive neural network control for switched nonlinear systems with unmodeled dynamics. (2nd October 2020)

• Record Type: Journal Article
• Authors: Long, Lijun; Bian, Yanan
• Is Part Of: International journal of robust and nonlinear control. Volume 30:Number 18(2020)
• Page Start: 7835
• Page End: 7856
• Publication Date: 2020-10-02
• Subjects: state constraints -- switched nonlinear systems -- unmodeled dynamics -- multiple Lyapunov functions -- average dwell time; Automatic control -- Periodicals; Control theory -- Periodicals; Nonlinear systems -- Periodicals; 629.836
• DOI: 10.1002/rnc.5113
• Languages: English
• ISSNs: 1049-8923

Abstract

Summary: This paper investigates the problem of state constraints‐based adaptive neural network tracking control for switched nonlinear systems with unmodeled dynamics. First, a switching‐dependent dynamic signal is designed to dominate unmodeled dynamics. Also, by introducing a nonlinear mapping, the limited state variables are transformed into new state variables without constraints in the switched systems framework, which permits removal of feasibility conditions for virtual control signals, when applying the barrier Lyapunov function (BLF) or integral BLF schemes. Second, adaptive neural network controllers of individual subsystems are constructed by exploiting backstepping and an improved average dwell time (ADT) method with dwell time reset, which guarantee that all signals in the resulting closed‐loop system are semi‐globally uniformly ultimately bounded under a class of switching signals with ADT and full state constraints are not violated, and furthermore, the tracking error converges to a small neighborhood of the origin. Finally, two examples, which include a two inverted pendulums as a practical example, are provided to demonstrate the applicability and effectiveness of the proposed design method.