Research On Ship Collision Avoidance Based On Firefly Algorithm

Due to the increasing number of ships at sea,ship collision accidents often occur,so it is very necessary to study ship collision avoidance.At present,the requirements of collision avoidance rules and steering range are seldom considered in ship collision avoidance,so the efficiency of collision avoidance is low.In this paper,the ship collision avoidance in open waters is studied.Because the consequences of ship collision accidents are very serious,there are high requirements for the efficiency of ship collision avoidance.The standard firefly algorithm is improved to enhance the convergence speed and accuracy of the algorithm.At the same time,the actual situation of ship collision avoidance is considered,and the randomly generated initial population is screened to improve the efficiency of collision avoidance.The collision avoidance rules are combined with the objective function of intelligent algorithm to make the collision avoidance path not only meet the requirements of economy and safety,but also meet the requirements of collision avoidance rules and steering range.(1)By reading the literature on ship collision avoidance,the basic knowledge of the general process of ship collision avoidance,ship field,and ship motion parameters is introduced.In order to choose a suitable collision avoidance algorithm,the advantages and disadvantages of genetic algorithm,particle swarm algorithm and standard firefly algorithm are analyzed,and the standard firefly algorithm is determined as the main algorithm for ship collision avoidance.(2)The principle,mathematical formula and algorithm flow of the standard firefly algorithm are introduced,and the algorithm is improved.When the firefly individual is far away,it is difficult to approach the excellent individual,so the position update formula is improved.Since the step factor of the algorithm is a fixed value and cannot be adjusted independently,an adaptive step size is proposed and the search space is constrained.Improved firefly algorithm improves the situation of falling into local optimum,and improves the speed and accuracy of convergence.(3)The improved firefly algorithm is applied to the research of ship collision avoidance.The collision avoidance path in the initial population that does not meet the actual situation is screened,and the calculation time of the algorithm for invalid individuals is reduced.A ship collision avoidance objective function based on safety,collision avoidance rules,steering range and economy is established,and the influence of various factors on collision avoidance is comprehensively considered by linearly weighting the objective function.At the same time,in order to solve the problem of smoothness and conform to the rotation of the ship,the Dubins curve is combined with the collision avoidance path.(4)Through MATLAB simulation,two ship encounter simulations are designed,which are three situations: head-on situation,crossing situation and overtaking situation,and the situation that the obstacle ship changes course in violation of the rules is simulated.The result proves that a feasible collision avoidance path can be found quickly,and the requirements of collision avoidance rules and steering amplitude can be met.The simulation of multi-ship encounters is designed and compared with the traditional real-time collision avoidance path planning.The collision avoidance method in this paper calls fewer algorithms and has higher collision avoidance efficiency.