Research On Path Planning And Tracking Control Of Autonomous Vehicle In U-Turn

With the development of autonomous driving technology,it can effectively solve the traffic accident rate.Vehicle U-turns are inevitable in daily driving,and most of the current research focuses on lane keeping,following control,lane change overtaking and automatic parking,so this paper focuses on vehicle U-turns on driving roads,and the main work is as follows:First,the real-life turnaround scenarios are analyzed and summarized,and three typical road turnaround scenarios are established to provide a map basis for path planning.The vehicle kinematics model,vehicle dynamics model and tire dynamics model are established,and the linear relationships between tire slip rate and tire longitudinal force,tire lateral deflection angle and tire lateral force are derived to provide a model basis for establishing path tracking controller.Secondly,for the problem that the vehicle cannot be effectively tracked by the path planned by the traditional artificial potential field,this paper proposes a motion constraint angle improvement method and a reversing judgment mechanism based on the vehicle kinematic model.Simulation results in three turnaround scenarios show that the improved artificial potential field method effectively avoids the problem that the algorithm falls into local optimum when planning the turnaround path,and the planned reference path meets the vehicle driving habits while satisfying traffic safety requirements.Then,this paper builds an MPC controller based on the vehicle transverse dynamics model,and establishes a comprehensive performance index optimization objective function with the minimum transverse error as the goal to realize the transverse tracking of the reference path by the self-driving vehicle.In this paper,a PID controller is built based on the vehicle longitudinal dynamics model to realize the longitudinal speed control of the self-driving vehicle.The simulation results show that the average absolute error of the MPC transverse controller in the transverse direction is 0.18 m in large curvature curves,while the average absolute error in the transverse direction is 0.69 m in the limit condition of continuous sharp turns;the average absolute errors of the PID controller in four different speed conditions are 0.06 m/s,0.02 m/s,0.03 m/s and 0.02 m/s.Therefore,the MPC lateral tracking controller and PID longitudinal tracking controller have high robustness.Finally,by establishing a joint CarSim/Simulink simulation model for different road scenarios,the results show that the vehicle can complete the turnaround task on both wide and narrow roads,and the average absolute errors of lateral tracking throughout the process are 0.12m and 0.26m,respectively,and the maximum errors during turning are 0.91m and 0.95m,respectively,and the average absolute errors of longitudinal speed When the self-driving vehicle makes a U-turn on a road with dynamic obstacles,it can effectively avoid and complete the U-turn task,but there is a risk of vehicle rear-end intrusion into the non-motorized lane and pressure line.In summary,the improved artificial potential field method proposed in this paper can provide reasonable and safe turnaround reference paths,and the established MPC lateral controller and PID longitudinal controller can realize the tracking of the reference path by the self-driving vehicle.Under this framework,the self-driving vehicle is basically able to complete the task of making U-turns under different road conditions.