车联网环境下车辆最优车速闭环快速模型预测控制

为解决智能交通系统中多辆车的最优车速预测问题，提高车辆燃油经济性及控制算法的执行效率，提出一种基于交通信号灯正时（SPAT）的方法来计算目标车速范围，避免车辆红灯停车；提出基于最优巡航车速和最优加速度跟随的模型预测控制建模方法；提出考虑效率反馈的闭环车速预测控制方法并采用快速模型预测求解最优目标车速。基于云计算的软件在环仿真结果表明，提出的控制方法可以避免车辆红灯怠速以及发生碰撞；相比于基准方法1和基准方法2，平均燃油经济性分别提高6.44%和37.1%；相比于基准方法1，系统平均驱动效率提高7.13%；快速模型预测程序执行效率为模型预测的13倍，且每一步长的实际计算时间小于计算步长，可以实现实时控制。

A Closed-loop FMPC of Optimal Velocities for Connected Vehicles

In order to solve the optimal velocity prediction problems of multiple vehicles in the intelligent transportation system, improve fuel economy of the vehicles and execution efficiency of the control algorithm, a signal phase and timing based method was presented to calculate the ranges of the target velocity thus red light stoppage might be avoided. A modeling method of model predictive control was put forward based on the tracking of the optimal cruising velocities and optimal accelerations. A closed-loop velocity predictive control incorporating with efficiency feedback was formulated and FMPC was adopted to evaluate the optimal velocity profiles. The cloud-computing based software-in-the-loop simulation results indicate that, based on the proposed method, red light idling and collision may be avoided. Compared with baseline method 1 and 2, the average fuel economy improves 6.44% and 37.1% respectively. Compared with baseline method 1, the average propulsion efficiency improves 7.13%. The execution efficiency of FMPC is 13 times higher than that of MPC, and real time control may be achieved since the operation time of a single step is shorter than the step size.