Research On Path Planning And Optimization Of Unmanned Ship Based On Improved Particle Swarm Optimization

When used in the path planning problem of USV,the standard particle swarm algorithm(PSO)has the characteristics of less adjustment parameters during planning and calculation,which is easy to operate and has good effect,but it is easy to fall into the local optimal solution,and the convergence is slow,generating The path is long.In addition,most of the previous studies only focus on a single optimization objective such as shortening the path length,which leads to poor performance of the planned path,and thus makes the unmanned ship perform poorly.This thesis studies the development status of unmanned ships at home and abroad,analyzes the previous research results,and compares the advantages and disadvantages of various commonly used unmanned ship algorithms.On this basis,two different improvement methods are proposed for the standard particle swarm algorithm: research Based on the improved particle swarm algorithm’s path planning capability for unmanned ships;and for the problem of single target existing in most path planning algorithms,the improved particle swarm algorithm’s unmanned ship path optimization capability is studied.An improved particle swarm optimization algorithm with genetic manipulation was proposed,which introduced the crossover mutation operation in the genetic algorithm in the improvement process to enrich the population diversity,reduce the path crossover and shorten the path length.At the same time,the inertia weight of the algorithm is adaptively adjusted to improve the convergence speed of the algorithm.Multiple experiments were carried out on the MATLAB simulation experimental platform with different numbers of sampling points.The results show that the improved particle swarm algorithm can effectively reduce the number of path crossings and shorten the path length.Compared with the standard PSO algorithm,the route quality is high and the convergence speed is fast.The path planning capability has been significantly improved.An improved particle swarm optimization algorithm based on hybrid simulated annealing is proposed.The algorithm introduces the initial temperature and cooling method of simulated annealing algorithm into the standard particle swarm optimization algorithm,and adds its ability to accept poor solutions with a certain probability.New hybrid algorithm to improve the ability of the new algorithm to jump out of local optima,thereby increasing the smoothness of the optimized path.Based on the initial path obtained by the Dijkstra algorithm,this thesis analyzes the optimization ability of the hybrid algorithm for the initial path.Using the MATLAB platform,several comparison experiments are carried out in two different environmental modeling scenarios.The experimental results show that: the hybrid simulation The improved particle swarm optimization of annealing is better than other algorithms in the application of USV path optimization.At the same time,the search path is shorter,the convergence speed is faster,and the convergence is stronger,which can effectively improve the global path optimization of USV in actual navigation.ability.Aiming at the problem that the original hybrid algorithm only considers the single objective of the shortest path,the improved particle swarm algorithm of hybrid simulated annealing is further optimized to improve its path optimization ability.Target optimization algorithm to fit the actual needs of unmanned ships.And a method to evaluate the performance of USV path optimization is proposed,so as to make a unified evaluation of the advantages and disadvantages of various USV path optimization algorithms.Using MATLAB simulation platform,the multi-objective algorithm is tested in two different environmental scenarios,and the comparison and analysis with the single-objective algorithm before improvement are carried out to verify the progress of the multi-objective improved particle swarm optimization in USV path optimization.