Collision avoidance path planning is still one of the essential problems in the design and application of an intelligent maritime navigation system. Its main obstacle is how to determine effective and cooperative collision avoidance maneuvers within a multi-ship encounter situation. By deconstructing a multi-ship encounter, this study adopted ship domain around target ships to assess the collision danger that own ship should avoid. Subsequently, the fitness function that has multiple dynamic obstacle constraints was designed in a two-dimensional map. Based on DE algorithm, a path-planning method was developed to compute collision-free and optimal navigation paths for ships. Simulation results show that the algorithm can generate a safe and suitable path from each perspective in a multi-ship encounter. The results also validate the practicality of the generated paths, consistency of the algorithm outputs and performance of the algorithm. It would be expected to provide a reference for collision avoidance decision making as well as contribute to the development of autonomous navigation systems.

避碰路径规划仍然是智能海上导航系统设计和应用中的基本问题之一。它的主要障碍是如何确定在多舰相遇情况下的有效和协作式避碰演习。通过解构多舰相遇，本研究采用目标舰船周围的舰船域来评估本船应避免的碰撞危险。随后，在二维图中设计了具有多个动态障碍约束的适应度函数。基于DE算法，开发了一种路径规划方法来计算船舶的无碰撞最优导航路径。仿真结果表明，该算法可以在多舰遭遇中从各个角度生成安全且合适的路径。结果还验证了所生成路径的实用性，算法输出的一致性以及算法的性能。有望为避免碰撞决策提供参考，并有助于自主导航系统的发展。