An Environment Adaptation Algorithm Based on RRT for Manipulator Path Planning

Abstract

Aiming at the problem of slow planning speed and path redundancy caused by repetitive searches of rapidly exploring random trees (RRT), an improved RRT algorithm (EA-RRT) based on environment adaptation strategy is proposed. The algorithm adaptively changes the step according to the vertex environment information around the nearest neighbor vertex to avoid repetitive searches in the same zone. Moreover, the target bias strategy is also introduced, which further accelerates the convergence speed of the algorithm. By comparing the performance of EA-RRT, RRT and RRT-star algorithms through MATLAB simulation and experiments, EA-RRT algorithm significantly improves the path planning speed and reduces path redundancy. Additionally, the effectiveness and practicability of EA-RRT algorithm are verified by robot operating system (ROS) in experiments. Both simulation and experiment have proved that EA-RRT algorithm is superior to the RRT algorithm in terms of operation time and path quality.