Scholarlink

<p>This paper proposes an event-triggered identification and control strategy for a class of nonlinear dynamical systems. We leverage online trained feedforward neural networks to approximate the plant dynamics. To improve efficiency, the network updates and control actions are triggered only when necessary based on the error evolution. This approach significantly reduces computational demands by utilising a small fraction of available data points. In addition, we demonstrate that this technique is far superior to traditional controllers like PD and PID in terms of the number of samples required with comparable identification and control performance when applied to a two-link robot manipulator.</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

Data-efficient Control of Two-link Robot Manipulator Using Event-triggered Neural Networks

  • 期刊论文

  • 发表日期：

    2025年8月13日

查看全文
  • Rashmi Ugarakhod,
  • Shikha Tripathi,
  • Koshy George

摘要

This paper proposes an event-triggered identification and control strategy for a class of nonlinear dynamical systems. We leverage online trained feedforward neural networks to approximate the plant dynamics. To improve efficiency, the network updates and control actions are triggered only when necessary based on the error evolution. This approach significantly reduces computational demands by utilising a small fraction of available data points. In addition, we demonstrate that this technique is far superior to traditional controllers like PD and PID in terms of the number of samples required with comparable identification and control performance when applied to a two-link robot manipulator.