Scholarlink

<p>With the increasing scale and complexity of multi-agent systems (MASs), developing control strategies for MASs with unknown dynamics is of significant importance. This paper thus investigates the problem of leader-following control for MASs subject to unknown dynamics and multiple faults. A neural observer with a structured bounded-parameter-layer neural network (NN) is proposed to observe immeasurable states and compensate uncertainties, replacing empirical basis functions in existing works. Based on the proposed neural observer, an iterative learning control (ILC) algorithm using only observation information is developed to handle random iteration lengths. A hierarchical control architecture integrating the observer and ILC layers resolves non-causal problems and mitigates transmission delays. Finally, simulations on intelligent connected vehicles (ICVs) confirm the effectiveness of the proposed scheme in eliminating unknown dynamics, multiple faults, and observation errors.</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

Neural Observer-Based Iterative Learning Leader-Following Control for Multi-Agent Systems Subject to Unknown Dynamics and Multiple Faults

  • 期刊论文

  • 发表日期：

    2026年1月26日

查看全文
  • Renyang You,
  • Quan Liu,
  • Shenghui Guo,
  • Zhiming Cui

摘要

With the increasing scale and complexity of multi-agent systems (MASs), developing control strategies for MASs with unknown dynamics is of significant importance. This paper thus investigates the problem of leader-following control for MASs subject to unknown dynamics and multiple faults. A neural observer with a structured bounded-parameter-layer neural network (NN) is proposed to observe immeasurable states and compensate uncertainties, replacing empirical basis functions in existing works. Based on the proposed neural observer, an iterative learning control (ILC) algorithm using only observation information is developed to handle random iteration lengths. A hierarchical control architecture integrating the observer and ILC layers resolves non-causal problems and mitigates transmission delays. Finally, simulations on intelligent connected vehicles (ICVs) confirm the effectiveness of the proposed scheme in eliminating unknown dynamics, multiple faults, and observation errors.