Research On Lateral Control Of Intelligent Vehicle Based On Adaptive MPC

Autonomous vehicle technology can improve the safety and convenience of automobile transportation.This technology has become a new direction for the development of intelligent vehicles in the future by solving the problem of low road traffic efficiency under urban conditions and ensuring driving safety.The purpose of this thesis is to optimize the proposed trajectory tracking control strategy in different ways and discuss its effectiveness in combination with the current status of control algorithms.In order to solve the problem of large tracking error between the reference trajectory and the real trajectory output by the planning system and the problem of vehicle driving stability,an adaptive controller is used as a solution to make the vehicle control strategy generalized at different speeds.The tracking error generated during the operation of the trajectory tracking controller is partly due to the inaccuracy of the vehicle model and the related controller parameters are fixed under different working conditions.Therefore,this thesis attaches great importance to the robustness and stability of the closed-loop system.With more and more scenarios of multi-input multi-output algorithm combined with automobile industry control,this thesis designs an adaptive trajectory tracking control algorithm based on vehicle dynamics model predictive control(Model Predict Control,MPC)to improve the accuracy of trajectory tracking and improve the driver’s ride comfort and vehicle stability.The lateral stiffness in the dynamic model is optimized online by the Forgetting Factor Least Square Algorithm(Forgetting Factor Least Square Algorithm,FFLSA),and the time domain parameters and weight coefficient matrix in the model predictive control are optimized online in real time by the rule-based control method.The control strategy with the best control effect can quickly reduce the tracking error when the vehicle trajectory changes greatly,ensure the tracking accuracy of the vehicle,and improve the comfort of the passengers.The results show that the adaptive time domain parameter controller can update the time domain parameters in real time with the change of vehicle speed,which improves the accuracy and stability of vehicle target tracking.The main research contents of this thesis are as follows:(1)The vehicle dynamics model is introduced and mathematically modeled.Finally,the simulation model is built based on Simulink,and the simulation results of the model are analyzed to verify its accuracy.(2)The vehicle parameter identification controller is built by Kalman filter theory and Dugoff tire model,and the gap between the estimated value and the true value of the cubature and unscented Kalman filter algorithm is compared.The results show that the lateral and longitudinal vehicle speed output by the unscented Kalman filter algorithm is more suitable as the input of the adaptive model predictive controller.(3)According to the principle of model predictive controller,an adaptive lateral trajectory tracking controller is designed.Then,the above three adaptive algorithms are built and deployed.The comparative analysis is carried out under the double lane change condition,and the real-time performance of the algorithm is verified.Finally,it is concluded that the MPC lateral controller based on time domain parameter adaptation has better control effect.Compared with the original MPC,the mean value of lateral acceleration decreases by 0.3 m/s~2,the mean value of lateral error decreases by 0.429 m,the mean value of sideslip angle decreases by 0.18 deg,and the mean value of yaw angle decreases by 0.44 deg.The average yaw rate increases by 0.14 deg/s,and the HIL test time fluctuates around 0.0106 s,which meets the real-time requirements.