基于扰动预测的网联车鲁棒协同巡航预测控制

Robust cooperative cruise predictive control for connectedvehicles based on disturbance prediction

针对车辆巡航过程中前车未来加速度不确定性引起的控制问题,提出了一种用于解决网联车巡航问题的鲁棒模型预测控制(Model predictive control,MPC)算法。采用车辆误差模型,将前车预测加速度与实际加速度之间的偏差视为加性扰动。根据已知的物理约束,先通过滚动预测的方法得到更紧凑的扰动多面体,然后对原系统的约束进行紧缩处理以降低传统tube MPC的保守性。基于传统tube MPC在控制器中引入前馈补偿,通过l2增益特性条件得到相应的线性增益,提出一种基于滚动扰动预测的tube MPC算法,并且证明了该算法是输入到状态稳定的。通过5辆车的数值仿真,验证了该算法的有效性。

In order to solve the control problem caused by the uncertainty of the future acceleration of the preceding vehicle during vehicle cruising,a robust model predictive control(MPC)algorithm for solving the cruising problem of connected vehicles is proposed.The vehicle relative error model is used,and the deviation between the predicted acceleration and the actual acceleration of the preceding vehicle is regarded as an additive disturbance.According to the known physical constraints of the vehicle,the disturbance polyhedron is predicted,and the rolling tightening constraints of the predicted disturbance polyhedron are used to reduce the conservatism of the traditional tube MPC.Based on the traditional tube MPC,the feedback term and disturbance compensation term are introduced into the controller.The corresponding linear gain part is obtained by l 2 gain property condition.A tube-based robust MPC algorithm via acceleration prediction is proposed for multi-constraint vehicle platoon control.It is proved that the proposed algorithm guarantees input-to-state stability.Finally,the effectiveness of algorithm is verified by a comparative experiment of five vehicles.