摘要
Abstract
In order to solve the problems of low convergence speed,unstable search and unsmooth path in the three-dimensional spatial path planning of the manipulator,an RRT* algorithm combined with dynamic constrained sampling was proposed.Firstly,the Gaussian sampling strategy of the dynamic target region is adopted,and the dynamic spherical subset is constrained by the sampling space with a certain probability to carry out Gaussian sampling,and the probability is dynamically adjusted by the number of collisions,so as to improve the path planning efficiency.In the process of reconnecting the parent node,QG-RRT*introduces a new cost function to replace the original Euclidean distance calculation,and adopts a forward culling strategy to remove the redundant part of the generation path.In order to meet the joint situation in the actual motion of the manipulator,the generated path was optimized by applying cubic B-spline curves for trajectory smoothing,and simulated in three-dimensional space and manipulator space by,and compared with RRT*,Bi-RRT*,and Inform-RRT*algorithms,the comparison results show that the improved algorithm is more efficient and has stronger plan-ning ability,and the improved algorithm uses ros for trajectory planning.关键词
机械臂/动态约束采样/代价函数/正向剔除/三次b样Key words
Dynamic constraint sampling/Adaptive step/Cost function/Forward pruning/Cubic B-spline分类
机械制造
