Scholarlink

<p>Optimized fuzzy prescribed performance control for stochastic networked nonlinear systems under denial-of-service (DoS) attacks is addressed in this research. Leveraging a meticulously crafted fuzzy estimator, unmeasurable system states during DoS attacks are modeled. Meanwhile, a simplified prescribed performance error transformation facilitates the derivation of a new Hamilton-Jacobi-Bellman equation, enabling the design of an optimized controller. Additionally, in the controller design procedure, fuzzy-logic systems integrated with reinforcement learning (RL) are utilized to approximate the unknown nonlinearities. In the optimized backstepping design, the proposed controller ensures the tracking error converges to the performance bound within a predefined finite time, even during DoS attacks. Moreover, employing Lyapunov stability theory, it is strictly proved that all signals in the closed-loop system are semiglobally uniformly ultimately bounded in probability. An event-triggered mechanism is introduced, which not only alleviates computational burden but also eliminates Zeno behavior. Ultimately, numerical and practical simulations are given to show the effectiveness of the proposed optimization method.</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

Adaptive event-triggered fuzzy algorithm for stochastic networked nonlinear systems under DoS attacks via reinforcement learning

  • 期刊论文

  • 发表日期：

    2026年5月6日

查看全文
  • Xin Zhang,
  • Wenjun Sun,
  • Junsheng Zhao,
  • Zong-Yao Sun,
  • Chaoxu Mu

摘要

Optimized fuzzy prescribed performance control for stochastic networked nonlinear systems under denial-of-service (DoS) attacks is addressed in this research. Leveraging a meticulously crafted fuzzy estimator, unmeasurable system states during DoS attacks are modeled. Meanwhile, a simplified prescribed performance error transformation facilitates the derivation of a new Hamilton-Jacobi-Bellman equation, enabling the design of an optimized controller. Additionally, in the controller design procedure, fuzzy-logic systems integrated with reinforcement learning (RL) are utilized to approximate the unknown nonlinearities. In the optimized backstepping design, the proposed controller ensures the tracking error converges to the performance bound within a predefined finite time, even during DoS attacks. Moreover, employing Lyapunov stability theory, it is strictly proved that all signals in the closed-loop system are semiglobally uniformly ultimately bounded in probability. An event-triggered mechanism is introduced, which not only alleviates computational burden but also eliminates Zeno behavior. Ultimately, numerical and practical simulations are given to show the effectiveness of the proposed optimization method.