Dynamic Task Assignment and Path Optimization for Multi-AUVs System

In scenarios such as volcanic eruption forecasting, seabed exploration, and subsea pipeline inspection, multi-autonomous underwater vehicle (AUVs) system is required to autonomously achieve close-range detection of target areas under mission constraints. This process relies on information exchange-based collaborative task planning, but disturbances from complex underwater environments and ocean currents directly affect the efficiency and safety of Multi-AUVs cooperation. A semi-closed ring bionic self-organizing map (SRBSOM) algorithm is proposed in this paper, which takes the mission urgency, currents and traveling salesman problem into consideration, and combines the hybrid bionic iterative network model, dynamic competitive assignment mechanism and semi-closed ring optimization to achieve dynamic task assignment for AUVs. Additionally, path post-processing and reorganization is performed through window filtering and corner smoothing, reducing control complexity and collision risks at turning points. Simulation results validate the environmental adaptability of the SRBSOM algorithm, which can provide a novel solution for the traffic planning of unmanned platforms in underwater environments.