An Improved Artificial Potential Field Method Based on DWA and Path Optimization

Abstract

The artificial potential field (APF) method is a simple and effective path planning approach. However, there is a fatal problem, which is that the robot can fall into local minima easily before reaching destination. Thus in this paper, we propose an improved APF method based on dynamic window approach (DWA) by evaluating points around robot in local minima with evaluation function, and choose the best point as next path point. Besides, we also propose an optimization algorithm to shorten the path. The main idea is to connect the consecutive points of the planned path while leaving enough secure space between robot and obstacles. The results show that our improved method can solve the problem of local minima and the optimization algorithm can plan a shorter path while consuming little computational time.