Dynamic path planning for ship collision avoidance based on adaptive raster map

As maritime traffic environments become increasingly complex, higher demands are placed on the safety and efficiency of path planning. To address the challenges of balancing computational efficiency and grid accuracy in environmental modeling, as well as the low compatibility and poor convergence of existing ship path planning algorithms with adaptive raster maps, this paper proposes an innovative approach. The method constructs an adaptive raster map by considering the distribution of static obstacles, collision risks, ship size, and maneuverability. Additionally, the ant colony optimization (ACO) algorithm is enhanced through the optimization of the heuristic function, initial pheromone distribution, and pheromone update rules. Based on this, the collision risk of the ship is used as the premise for local objective optimization to plan a safer and more efficient collision avoidance path. Experimental results demonstrate that, under the same conditions, the proposed method reduces memory usage by 92.3% compared to traditional methods, while achieving faster convergence and ensuring a balance between map modeling time and accuracy. The dynamically planned collision avoidance paths are also safer and more efficient.