Intelligent Collision Avoidance Of Surface Vechiles Under Multi-objective

With the development and utilization of marine environment by human beings,maritime traffic becomes more and more busy.Since the 1950 s,people began to consciously plan the rules of maritime traffic,people have realized that ship traffic safety is as important as other transportation.In maritime accidents,collision accidents account for more than half of all accidents.Nowadays,people equip ships with various collision avoidance aids,but collision accidents still emerge in endlessly.The fundamental reason is that although the driver gets more information,it is still the driver himself who makes the final decision of collision avoidance maneuver.However,in emergencies,especially in complex multi-objective surface environment,it is difficult for even experienced drivers to make correct judgments quickly.At this time,a system that can assist or even replace the driver in making collision avoidance decisions becomes particularly important.A new multi-objective collision avoidance strategy is proposed in this paper.Firstly,it introduces the relevant knowledge of the ship collision avoidance field and analyzes the basic methods of ship collision avoidance.In particular,it conducts a detailed analysis of the rules of maritime traffic and determines the ship's responsibility to avoid collisions with various encounters.For situations where multiple targets meet,separate different collision avoidance responsibilities.Second,the concept of collision risk is used to quantify the risk of collision between ships.And the shortcomings of the traditional BP neural network are improved,such as slow convergence,easy to fall into local minimum.The improved neural network was simulated by Matlab software,and its convergence speed and precision were compared with traditional BP neural network.Finally,it is a collision avoidance path planning algorithm based on distributed genetic algorithm.Using the risk estimation model established in the previous paper,combined with genetic algorithm,a set of intelligent collision avoidance methods for multiple targets is designed.When designing a genetic algorithm,only the factors considered are safety,economy,and compliance with maritime traffic rules.Based on this,the appropriate fitness function and objective function are designed.Finally,the program is designed in the development environment of Visual C++6.0.The program can set up at most three collision avoidance targets and can adjust the navigation parameters of the target ship.It is also possible to verify the convergence of the genetic algorithm under different genetic algorithm parameters.Finally,the whole system was simulated to verify the feasibility and reliability of the whole system.