Motion Path Planning of Sorting Robot Based on Extended RRT-Connect Algorithm

Abstract

Aiming at the problems of slow planning velocity, low efficiency and high path cost in the path planning of sorting robot by manual teaching or off-line planning, an improved rapid-exploration random tree algorithm is proposed to realize the autonomous obstacle avoidance movement of robot in the process of grasping in dynamic and unstructured environment. Based on the neighborhood radius of Voronoi visible area, the algorithm effectively controls the blind sampling near obstacles, improves the efficiency of effective sampling nodes. After introducing the idea of target gravity, the guidance of path generation and the operation efficiency of the algorithm are improved. For the motion path obtained by the planning algorithm, the Shortcut algorithm and cubic B-spline are used for optimization and smoothing, and finally the smooth motion path planning is realized. Through MATLAB two-dimensional static simulation and ROS (robot operating system) dynamic experiments, the results show that compared with the RRT algorithm and RRT-connect algorithm, the extended RRT-connect algorithm has been greatly optimized in path cost, search time and the number of sampling nodes, which proves the correctness, effectiveness and practicability of the proposed algorithm.