APF-RRT*: An Efficient Sampling-Based Path Planning Method with the Guidance of Artificial Potential Field

Path planning is a decisive module of mobile robots and its time efficiency significantly affects the safety of the robots. Sampling-based methods have achieved great success in the robotic path planning domain. However, poor time efficiency is still a serious limitation when they are applied to a crowded environment. In this paper, we combine the RRT* algorithm and artificial potential field(APF) technic and propose an efficient sampling-based path planning method named APF-RRT*. Utilizing the prior knowledge of the mission and the environment, we construct APFs for the start point, the goal point, the reference path, and the obstacles. Then we modify the random sampling step of the RRT* algorithm. With the guidance of APF, the random sample points are closer to the optimal path, and useless sample points greatly decrease. Results show that the proposed APF-RRT* outperforms state-of-the-art sampling-based methods in convergence rate, sampling effectiveness, and time efficiency.

I. Introduction

The mobile robot technic is beneficial for reducing labor and improving production efficiency, so it has become a hot research topic in recent years. Path planning is a key module of the mobile robot. The goal of this module is to generate a collision-free, smooth, and energy-friendly path in a complex environment. High time efficiency is the most important requirement to achieve this goal, so many studies focus on how to reduce the time cost of the path planning module. Despite that much progress has been made by previous research, path planning in a crowded environment is still a challenging job.