Re-planning method for unmanned surface vehicle path based on improved A* algorithm under dynamic sea conditions

The utilization of unmanned surface vehicles (USVs) has become increasingly prevalent in various activities, including ocean cruising, environmental monitoring, and autonomous navigation. As a result, the issue of safe and efficient path planning under complicated and dynamic sea conditions has attracted extensive research interest. By contributing to this growing body of research, this paper has proposed a dynamic re-planning method for unmanned ship paths based on the improved A* algorithm. To enhance the robustness and real-time performance of the algorithm for complex sea conditions, the dynamic re-planning method incorporates a multi-level environmental perception mechanism and adaptive cost modeling strategy. The first step of the proposed method, a spatiotemporal Gaussian Process regression model, is proposed to predict the velocity vector of local ocean current, assisting with the construction of a dynamic environmental prior field. The second step is the development of a risk perception cost function that incorporates the closeness to obstacles, changes in wave height, and predictive current drift to optimize path safety and energy consumption together. Subsequently, a dynamic re-planning framework with local-global fusion is developed, which uses a raster state map utilizing the underlying motion constraints, combined with a real-time velocity obstacle updating strategy to facilitate responses and maintenance of the path. The final part of the method outlines a strategy for search strategy, suggesting that a two-way priority queue structure can be incorporated to improve node expansion speed, collated with the directional guidance of heuristic functions and gradient based optimization techniques to balance search speed and path smoothness. The experimental results prove that the proposed method allows for a reduction in average path deviation by 23.7% in comparison to traditional average path in simulated ocean conditions testing.