Fast Path Planning for Polar Surface Unmanned Vessels Based on GI-ACO-A* Algorithm

Traditional A* algorithms exhibit two deficiencies in the realm of surface unmanned vessel path search: The first deficiency in path search is the excessive redundancy of visited nodes, leading to a high number of unnecessary node accesses that markedly degrades the efficiency of the path-finding process; The second deficiency is the neglect of the unmanned surface vessel's profile in the path planning. To enhance path search efficiency, the GI-ACO-A* algorithm (Goal-Induced A* under Ant Colony Optimization influence) is proposed, along with a navigation strategy integrating global path planning with local obstacle avoidance for unstructured environments. This approach features a global layer for overview path planning and a local layer for real-time obstacle evasion, adapting to the complex and dynamic surface conditions. Furthermore, this study processes radar imagery to obtain a sea ice distribution map and subsequently filters the radar image data to facilitate the modeling of a gridded environment. Simulation tests have validated its navigation safety and efficiency under complex surface conditions.