基于扩展雅可比矩阵的冗余液压驱动四足机器人运动控制

Motion control of a redundant hydraulic actuated quadruped robot based on extended Jacobian matrix

中文关键词  运动学冗余  逆运动学  摩擦锥  入地角  梯度投影法  扩展雅可比矩阵  四足机器人

英文关键词  kinematic redundancy  inverse kinematics  friction cone  landing angle  gradient projection method  extended Jacobian matrix method  quadruped robot

中文摘要

      针对冗余液压驱动四足机器人运动学逆解问题, 提出一种基于扩展雅可比矩阵的冗余液压驱动四足机器 人运动控制方法. 该方法既能解决冗余自由度带来的逆解多解问题, 还能使机器人足端入地角度满足摩擦锥要求 避免足端滑动. 首先, 规划机器人足端轨迹得到机器人足端速度, 在分析机器人足端入地角度对机器人运动性能影 响的基础上, 结合机器人腿部结构几何关系, 建立扩展雅可比矩阵, 确立机器人关节角度速度和足端速度的映射关 系, 即得到机器人关节角度的解. 然后, 在对角步态下, 通过仿真对传统的梯度投影法和提出的扩展雅可比矩阵法 进行对比, 理论分析及仿真表明传统的梯度投影法存在误差累积, 且在实时性上不如扩展雅可比矩阵法. 最后, 实 验验证了基于扩展雅可比矩阵逆运动学分析方法的可行性和有效性.

英文摘要

      To solve the inverse kinematics problem of redundant hydraulic actuated quadruped robots, an motion control method of a redundant hydraulic actuated quadruped robot based on extended jacobian matrix is proposed. This method can not only solve the multi solutions problem of inverse kinematics caused by redundant degrees of freedom, but also make the contact angle of robot foot meet the requirements of friction cone to avoid foot sliding. Firstly, the velocity of foot end effector of robot can be obtained from the planned foot trajectory. On the basis of analysis of influence of contact angle on locomotion performance of quadruped robot, the extended jacobian matrix method can be derived based on the geometric relationship of robot leg. Then, the mapping relationship between joints’ angular velocity and the velocity of the foot end effector is established. That’s the solution of joint angles of quadruped robot. Secondly, comparative simulations between the traditional gradient projection method and the proposed extended jacobian matrix method are implemented in trot gait. The theoretical analysis and simulation results show that there exists an error accumulation in the gradient projection method, and the proposed extended jacobian matrix method has a better performance in real-time. Finally, the feasibility and effectiveness of proposed method is validated by experiments.