Path Planning Based on Improved Hybrid A-Star Algorithm for USV

In response to the problems of long-term unreasonable regression and low search efficiency in the global path planning of Hybrid A-star algorithm, an improved global path search algorithm for Unmanned Surface Vehicles(USV) based on Hybrid A-star is proposed. This algorithm adopts a forward node expansion method and a segmented heuristic function calculation method. In the early stage of planning the path, the Euclidean distance calculation method is used, and the Dubins curve distance is used as the heuristic function when approaching the target point. At the same time, the USV physical model is considered in obstacle avoidance processing to ensure navigation safety during the planning process. The simulation results show that compared with the traditional Hybrid A-star algorithm, the improved algorithm reduces average planning time, planning path length, and node search quantity by $50 \%, 10 \%$, and 6 %, respectively. On the premise of ensuring the forward navigation of the USV, the search efficiency has been effectively improved.