Scholarlink

<p>This work investigates the implementation of distributed prescribed-time neural network (NN) control for nonlinear multiagent systems (MASs) using a dynamic memory event-triggered mechanism (DMETM). First, it introduces a composite learning technique in NN control. This method leverages the prediction error within the NN update law to enhance the accuracy of the unknown nonlinearity estimation. Subsequently, by introducing a time-varying transformation, the study establishes a distributed prescribed-time control algorithm. The notable feature of this algorithm is its ability to predetermine the convergence time independently of initial conditions or control parameters. Moreover, the DMETM is established to reduce the actuation frequency of the controller. Unlike the conventional memoryless dynamic event-triggered mechanism, the DMETM incorporates a memory term to further increase triggering intervals. Utilizing a distributed estimator for the leader, the DMETM-based NN prescribed-time controller is designed in a fully distributed manner, which guarantees that all signals in the closed-loop system remain bounded within the prescribed time. Finally, simulation results are presented to validate the effectiveness of the proposed 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

Neural-based prescribed-time consensus control for multiagent systems via dynamic memory event-triggered mechanism

  • 期刊论文

  • 发表日期：

    2025年2月14日

查看全文
  • Xiaohong Zheng,
  • Hui Ma,
  • Qi Zhou,
  • Hongyi Li

摘要

This work investigates the implementation of distributed prescribed-time neural network (NN) control for nonlinear multiagent systems (MASs) using a dynamic memory event-triggered mechanism (DMETM). First, it introduces a composite learning technique in NN control. This method leverages the prediction error within the NN update law to enhance the accuracy of the unknown nonlinearity estimation. Subsequently, by introducing a time-varying transformation, the study establishes a distributed prescribed-time control algorithm. The notable feature of this algorithm is its ability to predetermine the convergence time independently of initial conditions or control parameters. Moreover, the DMETM is established to reduce the actuation frequency of the controller. Unlike the conventional memoryless dynamic event-triggered mechanism, the DMETM incorporates a memory term to further increase triggering intervals. Utilizing a distributed estimator for the leader, the DMETM-based NN prescribed-time controller is designed in a fully distributed manner, which guarantees that all signals in the closed-loop system remain bounded within the prescribed time. Finally, simulation results are presented to validate the effectiveness of the proposed algorithm.