Scholarlink

<p>In this paper, a new push recovery control scheme is presented to counteract the adverse effects of external disturbances on locomotion stability, where the control strategies of one-step and multi-steps are incorporated. Under the newly developed control scheme, an efficient footstep planning approach and a variational-based stepping algorithm are proposed to achieve the robust push recovery of biped robots. To tackle the trade-off problem between locomotion stability and control performance, the step-size optimization problem is firstly formulated as a nonlinear programming (NLP) problem, and then an objective function is constructed to generate the appropriate next-step landing location under unexpected disturbances. Moreover, by designing an inverted-pendulum-model (IPM)-based state estimation method to detect center of mass (CoM) changes and estimate robot’s walking status in real time, the newly developed control scheme exhibits a good performance in maintaining locomotion stability. Finally, simulation and hardware experiments are carried out based on biped robots Jet-hr5 and Unitree-H1, respectively, which well illustrate the effectiveness of the proposed control scheme.</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

Online Energy-efficient Push Recovery Control for Biped Robot Using a Variational-based Stepping Approach

  • 期刊论文

  • 发表日期：

    2025年7月14日

查看全文
  • Zeming Liu,
  • Liang Yang,
  • Zhifeng Huang,
  • Guanyu Lai

摘要

In this paper, a new push recovery control scheme is presented to counteract the adverse effects of external disturbances on locomotion stability, where the control strategies of one-step and multi-steps are incorporated. Under the newly developed control scheme, an efficient footstep planning approach and a variational-based stepping algorithm are proposed to achieve the robust push recovery of biped robots. To tackle the trade-off problem between locomotion stability and control performance, the step-size optimization problem is firstly formulated as a nonlinear programming (NLP) problem, and then an objective function is constructed to generate the appropriate next-step landing location under unexpected disturbances. Moreover, by designing an inverted-pendulum-model (IPM)-based state estimation method to detect center of mass (CoM) changes and estimate robot’s walking status in real time, the newly developed control scheme exhibits a good performance in maintaining locomotion stability. Finally, simulation and hardware experiments are carried out based on biped robots Jet-hr5 and Unitree-H1, respectively, which well illustrate the effectiveness of the proposed control scheme.