Scholarlink

<p>This paper introduces an experience replay-based optimal event-triggered switched control (ETSC) strategy for the switched system with actuator failure. First, a fault observer is designed to estimate the error and compensate for the system input in real time. Next, a Lyapunov function is used to prove that the error of failure observation is ultimately uniformly bounded (UUB). Then, a dynamic event-triggered mechanism-based adaptive dynamic programming (ADP) is applied to solve the optimal ETSC problem. At each triggering instant, the controller determines which subsystem to switch and simultaneously updates the control input. The nominal cost function is approximated by designing a single-layer critic neural network (NN), and the experience replay technique is incorporated to update the NN weights. Stability analysis demonstrates that both the switched system and the NN weights are UUB. Finally, simulation results validate the effectiveness of the proposed ETSC scheme.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Experience replay-based dynamic event-triggered optimal control for nonlinear switched systems with actuator failure

  • 期刊论文

  • 发表日期：

    2025年2月22日

查看全文
  • Chong Liu,
  • Bingjie Yang,
  • Zhongxing Duan,
  • Weihua Li

摘要

This paper introduces an experience replay-based optimal event-triggered switched control (ETSC) strategy for the switched system with actuator failure. First, a fault observer is designed to estimate the error and compensate for the system input in real time. Next, a Lyapunov function is used to prove that the error of failure observation is ultimately uniformly bounded (UUB). Then, a dynamic event-triggered mechanism-based adaptive dynamic programming (ADP) is applied to solve the optimal ETSC problem. At each triggering instant, the controller determines which subsystem to switch and simultaneously updates the control input. The nominal cost function is approximated by designing a single-layer critic neural network (NN), and the experience replay technique is incorporated to update the NN weights. Stability analysis demonstrates that both the switched system and the NN weights are UUB. Finally, simulation results validate the effectiveness of the proposed ETSC scheme.