A Local Obstacle Avoidance Method for Unmanned Surface Vehicle Based on Acceleration and Angular Velocity

The traditional obstacle avoidance algorithm is based on position control. The ideal obstacle avoidance trajectory is derived by ship model. However, unmanned surface vehicle (USV) is an underactuated mission-driven system, and the cost of deviating from the predetermined path should be considered in the course of obstacle avoidance. Moreover, the actual motion path is often the result of the combined action of manual, automatic navigation and obstacle avoidance algorithms, which makes it difficult to obtain an accurate motion model. In this paper, the boundary between safety and danger is deduced from angle and distance, collision warning range is set, and new obstacle state parameters are defined. Then, the risk factors of different obstacles were quantified within the collision warning range, and the optimal course angle deviation and acceleration values were obtained, which were respectively output to the rudder angle controller and throttle controller. Simulation and real ship verification of the proposed algorithm show that the proposed method can effectively avoid obstacles and return to the preset track.