Research On Vehicle Path Planning Algorithm Based On GA Algorithm

With the continuous development of intelligent information processing technology,vehicles are applied in more and more fields,such as assembly production,transportation and manufacturing.Among them,path planning is one of the key technologies,which aims at how to generate the driving path of vehicles in real time and plan the route to obtain the optimal path under various dynamic environments.Traditional path planning algorithms usually assume that the environment is completely known and try to search for sharp turns and some polylines that contain the best path.Based on the actual project requirements,this thesis builds a simulation platform that satisfies the characteristics of the scene for the special traffic environment research.First,Discuss the method of vehicle path planning theoretically.After comparing the advantages and disadvantages of the current multi-path generation planning methods,the genetic algorithm is selected to solve the problem,and the vehicle path planning problem in complex environment is also an improved genetic algorithm based on bezier curve(BOBCIGA)is proposed.Secondly,In the path generation,bezier curve and segment path generation algorithm are introduced to make the path smooth.In the design of genetic operator,new smoothing,inserting,removing and optimizing the operator to compensate for the disadvantages of the basic operator,and using the self-adjusting strategy of crossover and mutation rate to optimize the algorithm accordingly.In order to solve the problem that the traditional genetic algorithm is prone to local optimal solution and long time,the two fitness matching functions are redefined.For the feasible path,the path length,smoothness,obstacle avoidance,objective function and other influencing factors are introduced.For the infeasible path,the influencing factors such as the length of the path itself,the collision depth of the moving vehicle and the obstacle,and the proportion of the unconventional line segment in the generated path are introduced.Finally,Under the scene modeling tool,three different levels of complexity are designed,simulation experiments are carried out,and the simulation results are analyzed.The path with the highest fitness value is output as the current optimal path on the average path length.The design path visual evaluation tool visually displays the algorithm results.The effects of different fitness parameters on the path planning resultsare discussed.Compared with other algorithms,the BOBCIGA algorithm proposed in this thesis saves at least 7% of the average search time compared with the traditional SGA algorithm;the average path length is reduced by at least 15%;compared with the A* algorithm,the average search time saves at least 10%;the average path length is reduced by at least 18%.In terms of path length and smoothness,the BOBCIGA algorithm proposed in this thesis is better than the two.In this thesis,a set of scenes for algorithm research is built by using the scene modeling tool.A lot of simulation experiments are carried out on the proposed algorithm,and all of them have achieved good results.The research and experimental results of this thesis show that the proposed BOBCIGA algorithm can solve the path planning problem of moving vehicles in complex environments.