Research On Obstacle Avoidance Problem Of Surface Unmanned Vehicle Based On Improved Genetic Algorithm

Water environment monitoring has been a global focus for a long time.In recent years,the development and protection of the oceans have been put on the agenda of all countries,exploring the water resources requires more manpower and material resources.The research of unmanned surface vehicle(USV)has attracted worldwide attention and has been applied in many fields,such as reconnaissance operation,ocean patrol,surveillance and search mission.At present,the marine environment and many reservoirs have been polluted seriously,so monitoring and collecting water samples is the most important thing,but fixed devices can only detect water quality parameters within a certain range,there are some factors to consider,such as application scope,installation device and so on.Portable devices need to be carried by people,sometimes even need to enter dangerous environment for monitoring and acquisition,there are hidden dangers to the personal safety of personnel.USV has the advantages of strong maneuverability,flexible direction control,low manpower consumption and high safety,however obstacles often exist on the surface of water to hinder the navigation of USV,so there is an urgent need for USV to be able to avoid collisions intelligently while driving,so as to complete tasks and reach the end point safely.This paper studies a hybrid algorithm based on improved genetic algorithm and improved artificial fish swarm algorithm,the algorithm can realize the safe navigation of USV in complex and multi-static obstacles water environment.The main work of this paper includes the following contents:Firstly,the genetic algorithm is improved,and the elite selection strategy is adopted to improve the computing speed and prevent losing the optimal solution.Aiming at the function of each operator in the algorithm,a protection operator and a elimination operator are added,and a small probability is set up,so that the optimal solution is retained and the number is increased generation by generation,the inferior solution is eliminated from generation to generation,at the same time the diversity of the population is guaranteed.Secondly,artificial fish swarm algorithm has two fixed parameters,visual field and step size.They cause low accuracy of the algorithm.Aiming at the disadvantage,exponential function is used to make the visual field and step size change adaptively with the iteration of the algorithm,so as to improve the accuracy of the algorithm and prevent local convergence.The improved genetic algorithm is combined with the adaptive artificial fish swarm algorithm,the genetic algorithm is inserted after the artificial fish swarm finishes its foraging,clustering and tailing behaviors.Protection,elimination and other operations are carried out to update the optimal solution on the bulletin board generation by generation,and to preserve the optimal path on the basis of improving the operational accuracy.Thirdly,the improved hybrid algorithm is compared with artificial potential field,ant colony optimization and particle swarm optimization,different paths are obtained under theenvironment of MATLAB,the comparison proves that the hybrid algorithm can avoid obstacles and has the best path,the practicability of the algorithm is verified.Finally,conclusions of this work are given,the deficiencies of this work and some valuable topics for future study are presented.