Research On Path Planning And Tracking Of Automatic Parking System Based On Genetic Algorithm

As an important part of advanced driving assistance,automatic parking system can not only improve the parking efficiency and reduce the driving burden,but also improve the safety and comfort of vehicles,improve the traffic environment and play an important role in the development of ITS.Path planning and tracking is an important part of automatic parking system,so it is of great significance to study the path planning and tracking of automatic parking system.Firstly,this paper analyzes the initial yaw angle of the existing automatic parking path planning and the scientific selection of the control parameters of the path tracking controller to determine the focus of this paper.Secondly,the vehicle model and parking environment model were simplified.The kinematics model of the automatic parking system was designed by ackermann steering principle.Then,taking the non-parallel initial state as the research object,the geometric relationship and related obstacle avoidance constraints of parallel parking and vertical parking path curves are analyzed,and the feasible starting area of parallel parking and vertical parking is established.Then,taking the minimum path length as the optimization objective,the multi beam non-linear trajectory equation is established to generate the minimum parallel and vertical parking path connecting the starting point and the ending point.Due to the problem of smooth and continuous curvature of the planning curve,so the β spline theory is used to fit and adjust the parking path curve,choosing the improved genetic algorithm to calculate and optimize the β spline,and the shortest and feasible parking path with continuous curvature,smooth curve and satisfying steering performance and obstacle avoidance constraints is obtained.Finally for the sake of accurate parking path tracking,path tracking controller was designed and analysis based on the LQR,because the selection of weighting matrix diagonal parameters on LQR controller is not scientific,so this paper selected the improved genetic algorithm to optimize the weighted matrix diagonal LQR controller parameters,and the results prove that the optimized LQR controller based on genetic algorithm responses faster to the target path,tracking to the target path more precisely and steadily by using the Matlab/Simulink with Carsim co-simulation.