控制理论(社会学)
机器人
扭矩
解耦(概率)
雅可比矩阵与行列式
模拟
计算机科学
地面反作用力
弹道
运动学
工程类
数学
控制工程
人工智能
物理
控制(管理)
经典力学
应用数学
天文
热力学
出处
期刊:Robotica
[Cambridge University Press]
日期:2022-01-26
卷期号:40 (8): 2855-2877
被引量:3
标识
DOI:10.1017/s0263574721001995
摘要
Abstract Currently, the body balance control algorithm of a quadruped robot executing trot gait motion is more complex and computationally intensive, which is not conducive to improving the real-time control performance of the robot. This paper proposes a six-dimensional spatial mechanics decoupling algorithm to enhance the balance control accuracy during trot gait while optimizing the computational effort. A 6 × 6 matrix is established to describe the relationship between six ground reaction forces of the diagonal supporting leg and six spatial forces and torques controlling robot degrees of freedom, which is optimized to reach the full rank. Decoupling calculation is adopted to obtain required ground reaction forces by matrix inverse operation, and forces are converted to joint motor torques utilizing the Jacobian matrix. The trajectory of the swing leg foot is generated based on cubic interpolation, and the robot motion speed is adjusted by selecting the landing point. This paper also proposes a region intervention control method based on center of mass projection to regulate the moving speed while ensuring the balance of the robot. Finally, the algorithm is verified by simulation using open source software Webots. The results show that when the robot moves at an average speed of 0.5 m/s, the lateral displacement change of the robot is less than 0.009 m, the height change is less than 0.003 m, and the rotating angles around the x, y, and z axes are less than 0.0036 rad, 0.0013 rad, and 0.001 rad, respectively.
科研通智能强力驱动
Strongly Powered by AbleSci AI