A dynamic velocity potential field method for multi-AUV cooperative hunting tasks

Multi-autonomous underwater vehicles (MAUVs) operate underwater with restricted information perception and swiftly altering environments. Existing methods face significant challenges in adapting to complex underwater environments. To solve this problem, a dynamic velocity potential field (DVPF) method is proposed to enhance the swarm performance for the cooperative hunting task of MAUVs in underwater environments with moving obstacles. The method is computationally affordable, applicable to many scenarios, and suitable for distributed control. First, a new attraction and repulsion model is designed by introducing velocity information. Without significantly increasing the computational cost, it makes AUVs more inclined to select short and safe paths. Then, for the path oscillation problem, the motion trajectory is smoothed by the velocity control module. Finally, through simulation results, the DVPF method is demonstrated to improve the performance of clusters in terms of task completion time, energy consumption ratio, and cluster path smoothness, and it is scalable for clusters with different velocities or in different environments.