Dynamic event-triggered-based adaptive fuzzy optimized control for slowly switched nonlinear system with intermittent state constrains

Abstract

This article presents a novel fuzzy event-triggered optimized strategy for slowly switched nonlinear system using reinforcement learning technique. For addressing intermittent state constraints problem, some improved shifting functions and barrier functions are designed in the adaptive backstepping process. Subsequently, by designing global performance functions with the discount term, the corresponding Hamilton-Jacobi-Bellman equation is derived and the optimal solution can be obtained under critic-actor structure. Different from the existing event-triggered methods, a novel dynamic event-triggered scheme for subsystem is developed, which not only saves communication resources, but also solves the mismatch behavior between controller and subsystem without restriction on maximum asynchronous intervals and the number of switch within a triggering interval. Furthermore, by using modified average cycle dwell time method and Lyapunov stability theorem, all signals of system are proved to be bounded. Eventually, numerical simulation results validate the superiority of the proposed approach, and this new scheme is applied to single-link robotic manipulator.