Trajectory Tracking Control Research For Autonomous Vehicle

The motion control of autonomous vehicle,a prerequisite for autonomous driving,is one of the core technologies of unmanned driving rese arch,taking an in-depth research into which is of great significance.This paper carried out the works mainly based on the requirements of stability and accuracy in the trajectory tracking control.The related works as follows.Firstly,the vehicle dynamics simulation model and the longitudinal control algorithm was established via PreScan and Matlab/Simulion,respectively.The longitudinal speed tracking control adopted a hierarchical control structure,wherein the upper controller is based on a linear time-varying model prediction control algorithm which obtains a desired longitudinal acceleration according to the desired longitudinal speed and the actual longitudinal speed in the driving.To control the vehicle dynamics of PreScan simulation model and achieve the purpose of longitudinal speed tracking control,the lower controller gained the ideal throttle angle or brake pressure based on the inverse longitudinal dynamics model.Secondly,the trajectory tracking control of autonomous vehicle was studied.The issue consisted of longitudinal speed tracking and lateral path tracking.Aiming at the accuracy of lateral path tracking control,an adaptive lateral control system based on Stanley control theory is proposed.Considering the lateral error,heading error and yaw angular velocity error between the vehicle actual position and the desired trajectory,the simulated annealing particle swarm optimization algorithm is used to optimize the controller parameters and obtained the optimal knowledge base based on the two-DOF vehicle kinematics model and the trajectory tracking error model.Also,the fuzzy control system was utilizing in achieving adaptive control effect under varied trajectories curvature and vehicle speeds.Still,the feasibility of the proposed control method was verified via the joint simulation environment of Matlab/Simulink and PreScan.Finally,the algorithm program for real vehicle experiment was written in the Qt integrated development environment under Ubuntu system.In order to verify the adaptability of the proposed control system to different vehicle speeds and different trajectories curvature,an offline map with various typical working conditions was designed.The trajectory file was expected to be combined with the control system by GPS/INS positioning navigator and offline map to further v alidate the effectiveness of the control system designed in this paper.