Research On Optimized Particle Swarm Optimization Algorithm For Unmanned Surface Vehicle Path Planning

In recent years,climate change,environmental anomalies,individual needs,and national defense have aroused widespread concern for the commercialization,scientific intelligence,and militarization of unmanned surface vehicles.Through multi-sensor fusion,unmanned surface vehicle(USV)has been effectively used in autonomous cruise,water quality sampling,garbage cleaning,environmental monitoring,marine rescue,target tracking and water quality monitoring and other marine fields,and because of the complicated sea conditions and target points more and different distribution density led unmanned surface vehicle routine completed sea homework time consuming energy dissipation,and even cannot effective implementation,so the path planning plays an important role in unmanned surface vehicle research.Heuristic algorithm is an effective method to solve the path planning problem and commonly used genetic algorithm,greedy algorithm and ant colony algorithm have complex execution mechanism and poor optimization performance,and the particle swarm algorithm for its convergence speed is fast,less set parameters,simple structure and easy to implement,the improvement can effectively realize unmanned surface vehicle of autonomous navigation and control.To design an efficient path for the multi-sensor integrated unmanned surface vehicle(USV),an improved hybrid particle swarm optimization(PSO)algorithm is proposed in this paper.The algorithm consists of four main stages.Firstly,the multi sub domain grouping strategy is adopted before the particle initialization,which reduces the complexity and calculation time of the path planning problem.To make full use of the global search advantage of the genetic algorithm,three mutation forms are used in the initialization stage.Besides,the multi-particle competition strategy based on permutation and the particle black box strategy based on greedy mechanism are used to iteratively search for optimization in the optimization stage,so that the algorithm can generate a local loop path effectively in each subdomain.After each subdomain path connection is completed,the local search method of reduced 4-opt is used to remove the path crossing.Based on the multi-sensor data,through the Monte Carlo simulation of 10 TSPLIB examples,the improved algorithm is compared with the existing algorithm,and the navigation application experiment of the self-developed unmanned ship is carried out.The experimental results show that the algorithm can achieve satisfactory results in terms of solution accuracy and computational efficiency.Firstly,this paper elaborates on the advantages,uses,and autonomous path planning of unmanned craft in detail.By studying various improvement measures and effects of particle swarm optimization algorithm,it summarizes hybrid optimization strategies,such as preprocessing of region decomposition,initial mutation population,multi-particle competition,greedy particle black box,and K-OPT.Secondly,the ideas and mechanisms of various heuristic algorithms,such as greedy algorithms and genetic algorithms,are deeply analyzed,and they are effectively introduced into particle swarm optimization algorithm.Through the mining analysis of C-means clustering and FCM clustering,the multi-subdomain clustering criterion based on the Cartesian coordinate system is summarized.Then,the improved particle swarm optimization algorithm was simulated by MATLAB for path planning,and through the box diagram,data table,path trajectory,dimensionless data and other forms,the performance of the algorithm is compared with my published relevant improved algorithm and other latest improved algorithms.Finally,the improved algorithm is applied to the self-developed navigation,guidance,and control(NGC)system of the unmanned surface vehicle,and the feasibility of path planning is verified by sea trials under several working conditions with different number and distribution characteristics of monitoring points.