This paper presents a real-time optimization strategy that targets on short-term energy consumption for the power-split hybrid electric vehicles (HEV) by focusing the mechanical motion of the powertrains. Instead of long-term energy optimization, which is usually investigated in the literatures on state of charge (SoC) management problem of HEVs, a short-term behavior of energy consumption is focused under the assumption that SoC of the battery is enough to provide the electric power required by the optimization. To this end, the transient mechanical motion of the powertrain is considered in the optimization problem instead of the battery SoC. The proposed strategy consists of two layers: the prediction of driver’s power demand based on the vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) communication and the optimization of energy consumption along the predicted power demand. To deal with the stochastic uncertainties in the driver’s power demand, Gaussian process regression model is developed for the prediction, and the optimization is formulated as a model predictive control problem with the mechanical model of the powertrain dynamics forced by the predicted driver’s demand. Finally, the simulation results are demonstrated where the driver’s demand is generated by a professional simulator under randomly eliminated traffic environment.

本文提出了一种实时优化策略，该策略通过关注动力总成的机械运动，针对动力分配混合动力电动汽车（HEV）的短期能耗。代替长期的能量优化（通常在混合动力汽车的充电状态（SoC）管理问题的文献中进行研究），在假设电池的SoC足以提供能量的前提下，重点关注能量的短期行为。优化所需的电力。为此，在优化问题中考虑了动力总成的瞬态机械运动，而不是电池SoC。提议的策略包括两层：基于车对车（V2V）和车对基础设施（V2I）通信的驾驶员功率需求预测，以及根据预测的功率需求进行的能耗优化。为了应对驾驶员动力需求中的随机不确定性，开发了高斯过程回归模型进行预测，并通过预测驾驶员需求推动的动力总成动力学机械模型将优化公式化为模型预测控制问题。最后，在随机消除的交通环境下，由专业模拟器生成驾驶员需求的仿真结果得到了证明。并将优化公式化为模型预测控制问题，并由预测驾驶员的需求推动动力总成动力学的机械模型。最后，在随机消除的交通环境下，由专业模拟器生成驾驶员需求的仿真结果得到了证明。并将优化公式化为模型预测控制问题，并由驾驶员预测的需求驱动动力总成动力学的机械模型。最后，在随机消除的交通环境下，由专业模拟器生成驾驶员需求的仿真结果得到了证明。