Abstract: This study took unmanned surface vessels (USVs) as the research objects to conduct a research on their global path planning. The traditional A∗ algorithm in global path planning for USVs has many shortcomings, such as the low efficiency of the algorithm, the length of the solution path being only grid optimal rather than actual optimal, and there being many turning points and uneven paths. Bidirectional parallelization and an elite extended search strategy were proposed to improve the search performance of the algorithm, and path pruning optimization and cubic B-spline curve optimization were used to break the constraints of the grid environment to further optimize the path. Matlab simulation experiments proved that the improved A∗ algorithm provided a more suitable and effective global path planning algorithm for USVs.

Key words: unmanned surface vessel (USV), path planning, A? algorithm, path optimization, elite extended search