Robust Predictive Control for Path Tracking of Intelligent Electric Vehicle Based on Wheel Corner Module
The wheel angle module integrated with the four-wheel independent drive/steering system can make the whole vehicle have excellent low-speed mobility and high-speed stability,and is an ideal carrier for intelligent vehicles,but too many angle/torque control inputs increase the difficulty of the whole vehicle control,and the uncertainty interference in actual driving also seriously affects the robustness of vehicle control,the common standard model predictive control cannot deal with these uncertainties.In order to improve the robustness of path tracking control for intelligent electric vehicle based on wheel corner module,a tube-based robust model predictive control(Tube-RMPC)method was proposed.The vehicle dynamics model and the path tracking model were built to analyze the uncertainty problems in the vehicle driving process and established the disturbance bounded linear time-varying uncertainty model system,constructs the nominal system model predictive control optimization problem,and proposes a robust invariant set calculation method that can be applied to the path tracking control system.A robust model predictive control strategy was constructed based on closed-loop system feedback.Based on MATLAB/Simulink and CarSim joint simulation and hardware-in-the-loop test platform,the effectiveness and real-time of the proposed control strategy were verified.The results show that,under the presence of uncertainties and disturbances from both the vehicle's own parameter uncertainties and external road surface adhesion coefficient uncertainties,the proposed Tube-RMPC reduces the maximum lateral displacement error by 30.2％and 48.4％,and the maximum lateral tilt angle by 31.6％and 7.8％,respectively,compared to the standard model predictive control.This effectively improves the vehicle tracking accuracy and stability.The proposed control strategy has good robustness while ensuring control accuracy,and has important reference value for the design of intelligent vehicle control system.
NETL
NSTL
万方数据
维普
