Research On Trajectory Tracking Control Method For Autonomous Vehicle

Autonomous vehicle has been regarded as the direction of the development of the future automobile industry,it receives widespread concern.Highly intelligent and automated self-driving car will free the drivers from the closed-loop control of "people-car-road",fundamentally alleviate the traffic congestion,and improve vehicle travel safety.Trajectory tracking control is one of the fundamental problems in the field of automatic driving research.It forms the basis of the realization of the automatic driving.The research mainly includes lateral tracking control and longitudinal tracking control.Because the vehicle has high nonlinearity and parameter uncertainty,it is of profound significance to design a longitudinal and lateral tracking control strategy which can overcome the above characteristics of the vehicle effectively.In view of the above problems,this study is mainly carried out as follows:This paper describes the system and the structure of the autonomous vehicle,and defines the tasks and functions of the self-driving car at each level.The three-degree-of-freedom dynamics model and the steering geometry model of the vehicle are established through appropriate assumptions and linear approximation,so as to provide the model basis of the vehicle for the longitudinal and lateral tracking control.Aiming at the problem of longitudinal tracking control,a pure pursuit control model,a synovial membrane control model and a model predictive control model are established based on the steering geometry model,the preview kinematic model and the vehicle dynamics model respectively.An emulation proof of the three kinds of established lateral tracking control models is conducted in the typical scene of double lane change.The results show that the tracking accuracy of the pure pursuit model is relatively poor under the condition of high-speed variable curvature road condition,but the controlled vehicle has good ride comfort under different working conditions.The synovial membrane control model has better tracking accuracy under different working conditions,but the ride comfort is relatively weak under high-speed condition.The model predictive control model shows both excellent tracking performance and ride comfort under different working conditions,but the real-time performance of the algorithm is relatively poor.In order to overcome the problem of longitudinal control,an upper controller of speed decision and a lower controller of vehicle speed tracking are designed respectively.The speed decision process of the upper controller takes into account the road curvature,the lateral tracking deviation and the current vehicle speed.Based on fuzzy PID,the lower controller of speed tracking is designed.A simulation verification is done for the designed longitudinal trading and control method under the simulation condition.The simulation results demonstrate that the upper control can predict the expected speed and the lower controller can control the vehicle to track the expected speed accurately.In terms of the integration of the lateral and the longitudinal control,based on the longitudinal tracking control and the lateral tracking control,considering the tracking effect of the lateral tracking control model under different working conditions,a switching strategy of the lateral tracking model is proposed.Through the simulation verification of the designed integrated tracking model under the simulation environment,the result shows that the design of the established integrated tracking control model can improve the ride comfort of the controlled vehicle while ensuring the tracking accuracy.