In view of the problems of multi-ship intersection, frequent dynamic disturbance and high collision risk in lock waters, this paper proposes a hierarchical path planning and real-time obstacle avoidance algorithm that combines multi-objective optimization and game theory. The improved A* algorithm is adopted in the global level, and a dynamic priority model is constructed by combining the ship size, speed and task urgency, so as to generate an initial path with both efficiency and fairness. By integrating Model Predictive Control (MPC) with the Dynamic Window Approach (DWA) at the local layer, the system achieves coordinated control for trajectory tracking and sudden obstacle avoidance, ensuring a safety distance of ≥2 times the vessel's length. At the same time, the extended Kalman filter (EKF) for LiDAR-AIS data fusion is designed to control the target prediction error within 0.5m and support 10Hz re-planning frequency. Conflict resolution adopts non-cooperative game model, and Nash equilibrium is solved by iterative optimal response algorithm to realize collision-free passage of multiple ships. Simulation results show that, compared with manual rule method and centralized MPC (CMPC), the minimum ship spacing of this method is increased by 73.5% in the three-ship intersection scene, and zero collision is achieved. In the continuous navigation of five ships, the transit time is shortened by 14.7%, the energy consumption is reduced by 15%, and the delay rate of high-priority tasks is reduced to 5%. The single-step calculation of the algorithm takes about 85ms, which meets the real-time requirements and has good robustness and engineering application prospects.
Xiao Hu (Sun,) studied this question.