A* Assisted Local Search for Mobile Robot Path Planning

초록

This paper proposes an A* assisted local search strategy to refine the paths found by evolutionary algorithms (EAs) for mobile robots. Specifically, this local search refines the paths from two aspects. First, for each path, all turning points are first identified. Then, for each sub-path connecting two adjacent turning points, the A* algorithm is executed to refine this sub-path. In this way, the sub-paths between any two adjacent turning points are optimized fully. Second, a redundant turning point elimination strategy is designed by removing unnecessary turning points to improve the smoothness of each path. In these two ways, both the path length and the number of turning points could be reduced largely. Particularly, this local search method can be embedded into any EA to solve the robot path planning problem. In this paper, it is embedded into two state-of-the-art ant colony optimization (ACO) algorithms for verification of its effectiveness. Experiments have been conducted on 15 environments of five scales ranging from 30×30 to 50×50 and three obstacle densities varying from 10% to 50%. The experimental results have revealed that this local search indeed helps the two ACO algorithms find much shorter global paths along with much fewer turning points.

키워드