Ship Autonomous Collision Avoidance For Complex Situation

Collision at sea is the significant issue affecting the safety of ship navigation which have caused huge economic loss and environmental pollution,especially in complex situation with high density of ships and large collision probability such as port and gulf.With the release of the rules of autonomous cargo ships by classification societies,researches on autonomous collision avoidance system attract wide attention.A deliberative collision avoidance trajectory planning algorithm was proposed in this paper focusing on trajectory planning and trajectory following in the autonomous collision avoidance system.Deliberative collision avoidance trajectory planning algorithm is based on A* search framework which redefines search space,improves state extension mode,redesigns state cost function,to satisfy maneuverability constraints,obstacles constraints and COLREGs-compliant constraints.The uncertainty of dynamic obstacle trajectory prediction is introduced to achieve the global trajectory probability optimal goal.In this paper,the trajectory planning and the trajectory following are coupled.The generation of the trajectory primitives and the following of the trajectory primitives adopt the LOS algorithm to follow the planned trajectory successfully.The parameters of the trimaran’s mathematical model is identified using the unscented Kalman filter.The autonomous collision avoidance system of a trimaran is constructed and hardware integration and software development are carried out.The effectiveness and superiority of the algorithms are verified through the simulation on trajectory planning and performance analysis of the trimaran in complex situations.The effectiveness and reliability of the autonomous collision avoidance system are verified by the autonomous collision avoidance navigation test,which has high academic and engineering value.