Collision-free Path Planning Based on Clustering

The goal of collision-free path planning is to find a continuous path from a starting position to goal position without any collisions. The collision-free path planning is a critical skill in mobile robot applications, and by now it is still a tough problem because mobile robots would easily run into the situation of planning for the current path while neglecting the global goal. In this paper, an approach of path planning based on clustering is presented, which combines the methods of both local collision-free path planning and global path intelligent search. This method firstly set a visual field for robot to imitate human being’s driving. Inside of the given visual field, the Maximal-Minimal Distance clustering rule is utilized to tackle the problem of current local obstacle avoidance. Then the strategy of searching an optimal path through a serial of navigating points is introduced with the consideration of a nearest path while searching global goal position. The simulation results prove the effectiveness of the proposed method.