Improved PRM for Path Planning in Narrow Passages

Abstract

In view of the shortcomings of Probabilistic Roadmaps (PRM) in the case of narrow passages, an improved method based on optimal sampling strategy is proposed. By sampling the dense area of the obstacles, the sampling points distributed inside the obstacle are selected and uniformly sampled by the distanced, so that the sampling point is generated around the obstacle in the free area, thereby increasing the number of sampling points in the narrow passages. The simulation results show that the improved PRM has more sampling points in the narrow passages than the standard PRM. And the time of path planning, the success rate and the path length are also significantly improved.