Hybrid path planning for USV with kinematic constraints and COLREGS based on improved APF-RRT and DWA

Intelligent algorithms are increasingly employed to optimize the ship path planning, ensuring safe navigation and preventing potential collisions. This paper proposes a hybrid path planning approach that incorporates kinematic constraints, which includes both global and local path planning. Firstly, the Artificial Potential Field (APF) is utilized to guide node growth, allowing the Rapidly-exploring Random Tree (RRT) to more efficiently navigate through multi-obstacle areas. To enhance the robustness of the algorithm, the gravitational coefficient optimization method and dynamic step-size strategy are introduced. Additionally, a novel RRT termination method is proposed to accelerate search speed by either using the node that satisfies the termination criteria as the last path point or by inserting a new node to fulfill all constraints. Furthermore, building upon the Dynamic Window Algorithm (DWA), a turning function based on maritime rules and the Collision Risk Index (CRI) are incorporated. New cost functions are established accordingly for different types of obstacles. The results from simulation experiments demonstrate the superior performance of the proposed method. The enhanced global path planning algorithm improves the efficiency of global path acquisition in diverse environments. Moreover, the proposed local path planning algorithm can accurately assess and respond to different scenarios and performs better than the traditional method in collision avoidance. •High density obstacle area is avoided with improved APF considering environment information.•Node surroundings considered dynamic step-size improves the RRT search efficiency.•A novel global search termination method enhances the search success rate.•A COLREGS-based turning function ensures safety, with CRI aiding collision avoidance.