Path Tracking For The Intelligent Vehicle Based On Model Predictive Control

Path tracking is one of the core technologies for intelligent vehicle to realize its intelligent behavior.Model predictive control has its unique advantages in dealing with the controlled system which is difficult to establish accurate mathematical model and has many constraints.Considering the complexity of vehicle model and the uncertainty of driving environment,this paper studies the path tracking of intelligent vehicle in different driving environment based on model predictive control.Firstly,the dynamic model and path tracking control method of intelligent vehicle are studied.Based on the tire model and vehicle dynamics model,an intelligent vehicle path tracking controller based on the model predictive control theory is built,and the optimal solution of the quadratic programming problem is studied in combination with the optimization objectives and constraints.The linear path condition and the doubleshift line condition are used as reference paths,respectively,and the designed path tracking controller is verified by the built MATLAB / Simulink and CarSim joint simulation platform.The simulation results show that the controller can complete the task of path tracking on the premise of ensuring the vehicle driving stability.The influence of vehicle speed on path tracking performance is analyzed.Secondly,based on the analysis of the influence of the angle constraint on vehicle path tracking performance,the adaptive control strategy of corner constraint is studied.By calculating the longitudinal force of tire under combined slip condition,combining with the theory of road adhesion coefficient and adhesion circle,the lateral adhesion of tire is obtained,so as to calculate the maximum steering angle of vehicle under the condition of no side slip.As the rotation angle constraint in the optimal solution process,the front wheel rotation angle increment is obtained,and the sum of the rotation angle increment and the control amount at the previous moment is used as the current control amount,thereby achieving continuous control of the steering wheel.The simulation results show that the tracking accuracy and driving stability of the designed control strategy are significantly better than that of the vehicle with fixed corner constraint at high speed or poor adhesion conditions.Thirdly,based on the analysis of the influence of sampling time and prediction horizon length on vehicle path tracking performance,the adaptive control strategy of the predicted horizon is studied.According to the national highway route design specifications,71 groups of driving conditions are divided with different speeds and road curvature radius,and the prediction horizon adaptive BP neural network is built based on the obtained optimal sampling time and prediction horizon length under different driving conditions.BP is used to predict the optimal sampling time and the prediction horizon length under different driving conditions.By inputting the predicted optimal sampling time and the prediction horizon length into the path tracking controller,the optimal vehicle control is achieved.The simulation results show that the prediction horizon adaptive vehicle can better take into account the path tracking accuracy and driving stability of vehicles at different speeds and different road curvature radius.Finally,the control strategy is verified by the hardware in loop simulation test of bramble pie 3B and MATLAB / Simulink.The results show that the tracking accuracy and driving stability of the two corresponding vehicles are similar.The calculation time of hardware in loop test in each cycle is slightly higher than that of dynamic simulation system.At the same time,it further explains the feasibility of raspberry pie as a hardware platform to build an intelligent vehicle path tracking controller.