多船确定性避碰行动最优决策模型

Multi-ship Deterministic Optimal Decision Making Model in Close-range Encounter under Variety Scenarios

针对多船复杂近距离会遇态势,提出了一种确定性避碰行动最优决策模型。考虑到船舶运动具有时变时滞特性以及风浪干扰的影响,结合神经网络和后推技术设计了船舶操纵运动控制器。同时结合多船会遇特点提出了多船避碰要素动态数学模型,并利用协同船舶领域确定符合安全阈值的最优避碰策略。当让路船驶过所有船的最小会遇距离点后,利用人工势场法规划让路船驶向目标点的最优路径。最后利用该模型对多船多场景近距离会遇态势进行仿真研究。仿真结果表明,所提出的模型能够在符合避碰规则的基础上生成安全可行的避碰策略,并有效减少避碰操纵次数,即通过单次避碰行动避让所有船舶。

In this paper a class of optimal decision-making model for multi-ship close-range collision avoidance is proposed.Considering the time-varying and time-delay characteristics of ship motion and the influence cause by wind and wave.An adaptive controller is designed by combining neural network and backstepping technology.On this basis,a multi-ship collision avoidance auxiliary decision-making model is constructed for multi-ship encounter situation.And the optimal collision avoidance strategy is determined by using the cooperate ship domain.While the give-way ship past the closest encounter distance points of all the target ships,the artificial potential field method is used to start planning the optimal path to the target points.Finally,the model is utilized to conduct simulation studies for multi-ship and multi-scenario close-range encounter situations.The simulation results shown that the proposed collision avoidance decision-making model,can effectively reduce the number of avoidance maneuvers,and avoid all the target ships through a single collision avoidance action.