Abstract: To improve the success rate and efficiency of the obstacle avoidance path planning of the cooperative robot in multiple obstacles environment,a collision detection method was proposed,and a low-oscillation-artificial-potential-field & Adaptive-Rapidly-exploring-Random-Tree (ARRT) hybrid algorithm was created.Low-oscillation-artificial-potential-field method was used to search,and ARRT was used for switching to escape with the conditions such as local minimum situation and collision until the target point was reached.In addition,to ensure the continuity of the joint angle between two continuously step,the shortest stroke algorithm was proposed to obtain the optimal inverse kinematic joint angle.To improve the quality of the planned path,a redundant path node deletion strategy was proposed and the path was fitted by four-time Bessel curve.Through simulation analysis,the manipulator had strong adaptability to adapt multiple obstacles environment,and the success rate of path search was higher than the traditional algorithm.The average time of path search was reduced from 26.1s to 3.6s compared with the classical RRT algorithm.Therefore,the algorithm search success rate and efficiency had been significantly improved.

Key words: collaborative robots, manipulator, obstacle avoidance path planning, low-oscillation-artificial-potential-field method, adaptive-rapidly-exploring-random-tree method

摘要： 为提高协作机器人在多障碍环境下的避障路径规划的成功率和效率,针对机械臂和障碍物提出碰撞检测方法,并提出低振荡人工势场—自适应快速扩展随机树(ARRT)混合算法进行路径规划,机械臂先采用低振荡人工势场法进行搜索,当遇到局部极小、碰撞等情况时切换成ARRT进行逃离,直至到达目标点。另外,为了在每个步长都取得最优的逆运动学关节角,保证前后步长对应关节角度值变化的连续性,提出最短行程逆解算法。为了提高规划后的路径质量,提出一种冗余路径节点删除策略,并使用四次贝塞尔曲线对路径进行拟合。经过仿真分析,机械臂在多障碍环境下对于环境复杂度的适应性强,路径搜索成功率高于经典算法,其平均路径搜索时间相比于经典RRT算法从26.1s下降到3.6s,算法搜索成功率和效率都得到显著改善。

关键词: 协作机器人, 机械臂, 避障路径规划, 低振荡人工势场法, 自适应快速扩展随机树法