Scholarlink

<p>For small-batch, highly diversified assembly scenarios, this paper proposes a digital twin-based hybrid potential field control method for human-robot collaborative assembly. By achieving millisecond-level synchronization of robot joint, workpiece, and operator data within a digital twin environment, the algorithm generates joint-level attractive/repulsive torques in real-time within a virtual-physical closed loop, thereby avoiding the stability issues caused by the inverse Jacobian matrix method. The constructed multi-objective attractive field can simultaneously satisfy multiple assembly constraints. By integrating joint limit constraints with an obstacle repulsive field, this method realizes collision avoidance, singularity posture evasion, and joint limit handling functions within a unified framework. Experiments conducted on the Unity-UR5e simulation platform demonstrate that, in five dynamic scenarios, the robotic arm can successfully track targets and safely avoid obstacles at an algorithm computation frequency of 10&#xa0;Hz, proving the method’s effectiveness and computational efficiency in high-dimensional real-time planning.</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

Composite potential field control for human-robot collaborative assembly robotic arm based on digital twin

  • 期刊论文

  • 发表日期：

    2026年1月27日

查看全文
  • Zhengjie Yang,
  • Liping Ma,
  • Hongjuan Yan,
  • Huayu Li

摘要

For small-batch, highly diversified assembly scenarios, this paper proposes a digital twin-based hybrid potential field control method for human-robot collaborative assembly. By achieving millisecond-level synchronization of robot joint, workpiece, and operator data within a digital twin environment, the algorithm generates joint-level attractive/repulsive torques in real-time within a virtual-physical closed loop, thereby avoiding the stability issues caused by the inverse Jacobian matrix method. The constructed multi-objective attractive field can simultaneously satisfy multiple assembly constraints. By integrating joint limit constraints with an obstacle repulsive field, this method realizes collision avoidance, singularity posture evasion, and joint limit handling functions within a unified framework. Experiments conducted on the Unity-UR5e simulation platform demonstrate that, in five dynamic scenarios, the robotic arm can successfully track targets and safely avoid obstacles at an algorithm computation frequency of 10 Hz, proving the method’s effectiveness and computational efficiency in high-dimensional real-time planning.