融合时延和参数不确定性的智能汽车轨迹跟踪鲁棒协同控制

To solve the degradation problems existing in vehicle trajectory tracking control due to the parameter uncertainty and network delay, a predictive control strategy was proposed for output feedback robust model based on fast convergence of state error. Firstly, considering the uncertainty sources of vehicle parameters and the random delay characteristics of network signals, a multi-center control model was constructed with the incorporation of two type uncertain terms. Then, a robust observer was taken to observe the state variables accurately, and a predictive controller with fast constraints was designed effectively for the robust model to improve trajectory tracking accuracy and yaw stability of the vehicles. The test results show that the proposed strategy can effectively eliminate the impact of two type uncertainties on tracking control. Compared with the traditional linear model predictive control and robust model predictive control methods, the trajectory tracking accuracy of the new proposed robust model can be improved by 83.70% and 19.41% respectively, and Yaw stability can be improved by 72.88% and 40.74% respectively. The actual vehicle tests show that the proposed strategy can provide a better real-time performance and feasibility.