Robust AI algorithm:a key technology to optimize ship intelligent collision avoidance system

The current ship artificial intelligence anti-collision system is weak robust,which is characterized by displaying all detection objects,on the one hand,it will cause a load burden on the system,and on the other hand,the pilot needs to manually extract objects with collision risk and make collision avoidance decisions.In this paper,which is based on fault tree is used to calculate the collision probability in combination with industry experience in collision avoidance of innovation methods of artificial intelligence training platform,for the first time put forward to improve artificial intelligent collision avoidance system automatically choose to have strong robustness,collision danger mark makes the system automatically show key target of risk of collision,help the driver to make decision-making of collision avoidance.