Scholarlink

This research focuses on implementing Machine Learning Impedance Control for robot arms to ensure both effective and safe interaction within unknown environments. The designed Machine Learning system comprises two key steps. Firstly, a Neural Network Classifier framework categorizes attributes by assigning labels to various impedance behaviors. Secondly, employing a regularization technique, the system determines the appropriate impedance behavior as a command signal for the internal control loop. Utilizing a backstepping-based algorithm alongside Taylor’s parameter adaptation within the control loop facilitates achieving effective interaction. To validate the efficacy of this control approach, the proposed method undergoes testing and simulation on a two-link robot manipulator across varying environmental stiffness levels.

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

Robust Learning-Based Impedance Control of Robotic Manipulators in Unknown Environments

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • G.  Reza Nazmara,
  • Pedro Neto,
  • A. Pedro Aguiar

摘要

This research focuses on implementing Machine Learning Impedance Control for robot arms to ensure both effective and safe interaction within unknown environments. The designed Machine Learning system comprises two key steps. Firstly, a Neural Network Classifier framework categorizes attributes by assigning labels to various impedance behaviors. Secondly, employing a regularization technique, the system determines the appropriate impedance behavior as a command signal for the internal control loop. Utilizing a backstepping-based algorithm alongside Taylor’s parameter adaptation within the control loop facilitates achieving effective interaction. To validate the efficacy of this control approach, the proposed method undergoes testing and simulation on a two-link robot manipulator across varying environmental stiffness levels.