Adaptive Path Planning for a USV: Enhanced RRT and DWA Based on Multi-Sensor Fusion Slam
Fuyu Zhang, Ning Wang, Shumin Fan · 2025
The traditional path planning algorithm can not make full use of the environmental information, and the path can not guide the unmanned surface vehicle (USV) movement well. Addressing the challenge posed by the inability of electronic chart system to provide comprehensive obstacle information for path planning, this paper adopts multi-sensor fusion simultaneous localization and mapping (SLAM) to express information about obstacles in navigable areas and build a point cloud map. The adaptive step size strategy and the target attraction sampling strategy are adopted to enhance rapidly-exploring random tree (RRT) as adaptive-rapidly-exploring random tree (A-RRT). The path is optimized to align with the unique movement characteristics of the USV better. For dynamic obstacles on the global path, the evaluation function of the dynamic window approach (DWA) is designed considering the disturbance from winds and waves to the USV. The dynamic adjustment strategy of the weight of the evaluation function for different marine environments such as wide areas and narrow waterways is adopted, and the USV employs various avoidance strategies according to international regulations for preventing collisions at sea. The simulation results show that the proposed path planning algorithm can realize the safe and efficient avoidance of static and dynamic obstacles.