Scholarlink

<p>This paper proposes a coordinated control method based on collision adaptive detection for constrained bipedal reconfigurable robots (BRRs) facing external constraints. Firstly, a dynamic model for BRRs is established by considering both kinematic and dynamic uncertainties. The reference plantar force is obtained via the quadratic programming strategy to maintain the standing balance of BRRs. Secondly, considering the external collision in the extreme environment, a model-free adaptive collision detection technique is presented. It corrects discrepancies between the approximated and actual torques in the robotic dynamics model, enabling accurate detection of collisions. Then, a fusion error function is designed reflecting the plantar force and joint position of the BRRs. The model-free coordinated control strategy of the BRRs is obtained under the radial basis function neural network (RBFNN)-based identifier. Through Lyapunov stability theory, the tracking error of closed-loop BRRs system is proved to be asymptotically stable. Finally, the effectiveness of the proposed method is demonstrated by a bipedal reconfigurable robot platform.</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

Collision adaptive detection-based model-free gait coordination control for bipedal reconfigurable robots

  • 期刊论文

  • 发表日期：

    2025年8月13日

查看全文
  • Yuanchun Li,
  • Meng Qu,
  • Qiang Pan,
  • Yuan Yao,
  • Tianjiao An,
  • Bing Ma

摘要

This paper proposes a coordinated control method based on collision adaptive detection for constrained bipedal reconfigurable robots (BRRs) facing external constraints. Firstly, a dynamic model for BRRs is established by considering both kinematic and dynamic uncertainties. The reference plantar force is obtained via the quadratic programming strategy to maintain the standing balance of BRRs. Secondly, considering the external collision in the extreme environment, a model-free adaptive collision detection technique is presented. It corrects discrepancies between the approximated and actual torques in the robotic dynamics model, enabling accurate detection of collisions. Then, a fusion error function is designed reflecting the plantar force and joint position of the BRRs. The model-free coordinated control strategy of the BRRs is obtained under the radial basis function neural network (RBFNN)-based identifier. Through Lyapunov stability theory, the tracking error of closed-loop BRRs system is proved to be asymptotically stable. Finally, the effectiveness of the proposed method is demonstrated by a bipedal reconfigurable robot platform.