Path Planning for Multipoint Seabed Survey Mission Using Autonomous Underwater Vehicle

Following an initial survey, marine geologists typically designate discrete targets requiring further observation or sampling; such points may be distributed over large sea areas. Advances in the endurance of autonomous underwater vehicles (AUVs) make it feasible to survey multiple targets in a single dive, and multi-AUV cooperation can further improve efficiency, thereby motivating voyage-level global path planning. Prior studies have largely addressed local coverage around individual targets or point-to-point routing in the presence of currents and obstacles, while large-scale multipoint missions remain comparatively underexplored. Classical assignment frameworks from problems such as school bus routing are likewise difficult to apply directly, because every target must be visited exactly.

This paper decomposes the problem into two stages by a greedy strategy. Iterative K-means clustering first determines the layout of support-vessel anchor points; an ant-cycle system then optimises the visit order from each anchor to neighbouring targets. Simulation results confirm the effectiveness of the algorithm and indicate its suitability for resource-constrained multipoint seabed surveys and multi-AUV deployments.