Dynamic obstacle-avoiding motion planning for manipulator based on improved artificial potential filed

DOI编号  10.7641/CTA.2018.70187
2018,35(9):1239-1249

 作者 单位 邮编 谢龙 浙江大学控制科学与工程学院 310027 刘山 浙江大学控制科学与工程学院 310027

本文针对刚性机械臂提出了一种基于改进势场法的动态避障规划算法. 势场产生只作用于机械臂末端的 引力和作用于机械臂与障碍物最近点的排斥力, 并依据动力学定律分别产生吸引速度和排斥速度. 本文直接在笛卡 尔空间中构造基于目标速度的吸引速度, 使机械臂能够追踪动态目标; 构造方向依据障碍物速度变化的排斥速度, 使机械臂连杆也具备避开动态障碍的能力. 最终两种速度在机械臂关节空间合成, 为机械臂规划一条无碰路径并 控制机械臂追踪动态目标. 仿真结果证明了该方法的有效性.

This paper proposed a dynamic obstacle-avoiding motion planning approach based on the improved artificial potential field for a rigid manipulator. The attractive potential field generates an attractive force which only affects the end of the manipulator. While the repulsive potential field generates repulsive force which affects the manipulator through the closest point between manipulator and obstacle. According to dynamics, these two kinds of forces generate attractive velocity and repulsive velocity respectively. In this paper, attractive and repulsive velocities were contructed in the cartesian space directly. The attractive velocity based upon the velocity of goal allows the manipulator to track a moving goal, while the repulsive velocity based on the change of obstacle’s velocity enables the manipulator to avoid moving obstacle. The attractive and repulsive velocities together plan a collision free path for manipulator in joint space, and control it to keep track of a moving goal. Simulation results are provided under some representative circumstances to show the effectiveness of the proposed method.