A two-stage dynamic collision avoidance algorithm for unmanned surface vehicles based on field theory and COLREGs

This studies the dynamic collision avoidance (CA) problem of unmanned surface vehicle (USV) under the International Regulations for Preventing Collisions at Sea (COLREGs, short for Collision Regulations). A more realistic virtual spatial velocity field and two-stage CA area are proposed to reflect the movement of the target ship (TS), and the heading state of the USV is divided into a safe state and an emergency state. A decision model of ship steering for CA by analyzing the obligatory behavior of giving way to ships in the COLREGs is proposed. First, in the safe state, a novel of CA algorithm is proposed by using particle swarm optimization (PSO) and the problems of avoiding the TS and sailing to the destination are considered simultaneously. Second, in the emergency state, a new CA algorithm is developed by combining the decision model of ship steering for CA with the artificial potential field (APF) method, when considering avoiding the TS. Finally, the effectiveness of the algorithms are proved by simulation.

Automated summary

This study investigates the dynamic collision avoidance problem of unmanned surface vehicles (USVs) under the International Regulations for Preventing Collisions at Sea (COLREGs). A more accurate virtual spatial velocity field and a two-stage collision avoidance area are proposed to reflect the movement of the target ship (TS). The heading state of the USV is categorized into a safe state and an emergency state. A decision model of ship steering for collision avoidance is presented by analyzing the obligatory behavior specified in the COLREGs. Two algorithms using particle swarm optimization and artificial potential field method are developed for safe state and emergency state, respectively. Simulation results demonstrate the effectiveness of the algorithms.