Dynamic Path Planning Algorithm for Unmanned Surface Vehicle Under Island-Reef Environment

The marine environment surrounding islands and reefs exhibit significant time-varying characteristics, which require the path planning algorithm of an unmanned surface vehicle (USV) to have good environmental adaptability. To this end, a dynamic path planning algorithm for USV under an island-reef environment is proposed. The algorithm fully considers these environmental features, such as ocean currents, tides, and winds, to construct an environment model. The environment model is combined with the USV motion model to construct a path planning model under the constraint of environmental disturbance time windows. This approach ensures that the path planning model can adapt to the time-varying marine environment. At the same time, to improve the safety of USV navigation by effectively predicting and timely avoiding passing ships, the velocity obstacle method is introduced for establishing a collision risk assessment model that detects the threat level of passing ships in the vicinity of the USV. Based on the decision-making basis generated by the model, a path replanning model is constructed under the constraint of collision detection time windows. This improves the dynamic collision avoidance capability of the path replanning model. Furthermore, to speed up the solution speed and accuracy of the path planning model and the path replanning model, the multiobjective particle swarm optimization algorithm is improved in three aspects: particle coding method, particle update strategy, and external archive maintenance mechanism. Simulation results show that the algorithm can enable USV to safely avoid multiple close-range dynamic obstacles under maritime rules, while also demonstrating a high level of adaptability to the time-varying marine environment.