Path planning of unmanned surface vehicle based on improved A* algorithm

This paper presents an in-depth optimization of the shortcomings inherent in the traditional A* algorithm for path planning of Unmanned Surface Vehicles (USVs). Special attention is given to addressing problems such as ship width, redundant nodes and frequent steering encountered during the actual navigation process of the USV. Firstly, through dilation processing of obstacles, the influence area of the obstacle's edge expands, efficiently circumventing the crossing point between the planning route and the obstacle, thus improving the accuracy of path planning algorithm. Secondly, the collinear path nodes are streamlined, retaining only the head and tail path nodes, thereby lessening the complexity and energy consumption of the path. Finally, the strategy of backtracking path nodes is utilized to optimize the path by eliminating unnecessary nodes, compressing the path length and significantly decreasing the quantity of turns. This approach further enhances the efficiency of path planning and the smoothness of navigation. Through experimental simulation, the improved A* algorithm proposed in this study demonstrates a better planning effect contrasted with the traditional A* algorithm on maps of different sizes. In actual waterway maps with dimensions of 100*100, the path length is reduced by 4.4 percent, the quantity of turns decreases by 78.9 percent and the number of navigation nodes decreases by 94.6 percent, which effectively enhances navigation efficiency and safety.