Path Planning for Mobile Robot Based on an Improved Probabilistic Roadmap Method

Probabilistic roadmap method planners have been work in path planning ofmobile robots, but sampling narrow passages in robot configuration spaceremains a challenge for PRM planning. This paper presents an improvedprobabilistic roadmap method for finding paths through narrow passages.A key ingredient of the new method is Branching random walk, which ismore integrated into probabilistic roadmap to capture the connectivityof free spaces with difficult narrow passages. The paper implementedthe planner and tested it on articulated robots in 2-D environments.Simulation shows that the method enables relatively small roadmaps toreliably capture the connectivity of configuration spaces withdifficult narrow passages. The method adopts to search path by use ofmore integrated of probabilistic roadmap, decreases time of collisiondetection, gets local optimal path, and improves the efficiency ofalgorithm.