Scholarlink

<p>This paper investigates the adaptive learning secure consensus problem for a class of nonlinear multi-agent systems (MAS) subject to Denial-of-Service (DoS) attacks. To overcome the issue of repeated differentiation of discontinuous switching signals induced by DoS attacks during the backstepping design, a novel high-gain switching observer is proposed to effectively estimate the leader’s state information. Furthermore, an incremental adaptive parameter update mechanism is developed to significantly reduce computational complexity. Notably, the proposed neural-network-based adaptive secure control protocol successfully circumvents the numerical integration issues inherent in traditional integral adaptive methods while incorporating robustness analysis into the Backstepping framework, thus simplifying the controller structure and lowering implementation complexity. Theoretical analysis demonstrates that the distributed protocol ensures the uniform boundedness of all closed-loop signals while preserving system tracking performance. Finally, numerical simulations verify the efficacy of the proposed adaptive learning control algorithm.</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

Observer-based incremental adaptive learning control for nonlinear multi-agent systems against DoS attacks

  • 期刊论文

  • 发表日期：

    2026年5月26日

查看全文
  • Shuting Wang,
  • Jinsha Li,
  • Junmin Li

摘要

This paper investigates the adaptive learning secure consensus problem for a class of nonlinear multi-agent systems (MAS) subject to Denial-of-Service (DoS) attacks. To overcome the issue of repeated differentiation of discontinuous switching signals induced by DoS attacks during the backstepping design, a novel high-gain switching observer is proposed to effectively estimate the leader’s state information. Furthermore, an incremental adaptive parameter update mechanism is developed to significantly reduce computational complexity. Notably, the proposed neural-network-based adaptive secure control protocol successfully circumvents the numerical integration issues inherent in traditional integral adaptive methods while incorporating robustness analysis into the Backstepping framework, thus simplifying the controller structure and lowering implementation complexity. Theoretical analysis demonstrates that the distributed protocol ensures the uniform boundedness of all closed-loop signals while preserving system tracking performance. Finally, numerical simulations verify the efficacy of the proposed adaptive learning control algorithm.