Geometric Collision Avoidance Algorithm for Unmanned Surface Vehicle Based on Multi-objective

Aiming at the problem of autonomous collision avoidance when multiple unmanned surface vehicles (USV) meet, and considering the possible existence of abnormally traveling USVs, this paper proposes a distributed collision avoidance algorithm including collision risk detection based on geometric collision avoidance algorithm. First, introduce obstacles Object prediction trajectory and weight factor improve the distance evaluation function of traditional geometric algorithms, and improve the ability of USV to avoid multiple dynamic obstacles. At the same time, a new rule evaluation function is introduced to constrain the USV's avoidance action. Then, the expected speed and heading are improved. The existing collision risk detection algorithm is used to reduce the trajectory fluctuation caused by the change of collision risk. Then, in the process of arriving at the target to distinguish the waypoint, the pure proportional guidance method ensures the safety of the unmanned boat while avoiding intruders. Finally, the simulation verification is carried out in the multi-static and dynamic mixed obstacle scene. The results show that even if there are various dynamic and static obstacles, the geometric collision avoidance algorithm still guarantees the normal USV safe avoidance action.