Path planning for unmanned surface vehicle based on predictive artificial potential field

Abstract

The path planning for high-speed unmanned surface vehicle raises more complicated requirements to reduce sailing time and save energy. In this article, a new predictive artificial potential field is proposed using time information and predictive potential to plan a smoother path. The principle of artificial potential field with vehicle dynamics and reachability in local minimum is studied. According to global and local path planning, the most state-of-the-art traditional artificial potential field and its drawback are analysed at first. Then we proposed predictive artificial potential field with three modifications: angle limit, velocity adjustment and predictive potential to improve the feasibility and flatness of the generated path. In addition, we compare the performance between traditional artificial potential field and predictive artificial potential field, where predictive artificial potential field successfully restricts the maximum turning angle, cuts short sailing time and intelligently avoids obstacle. From the simulation results, we also verify that predictive artificial potential field can solve concave local minimum problem and enhance the reachability in special scenario. Therefore, the more reasonable path generated by predictive artificial potential field reduces sailing time and helps conserve more energy for unmanned surface vehicle.