Power Demand Prediction-based Real-Time Optimal Energy Management Strategy for HEVs

Abstract

This paper presents a real-time optimal energy management strategy with prediction of driver’s power demand for parallel HEV. By modeling a driver’s behavior and identifying the parameters of the model at each step, the power demand is predicted for several seconds in real time. Optimal energy management problem is solved by MPC using the predicted data. Simulations over various driving routes were conducted to confirm the driver’s behavior model can contribute to the better solution of the proposed optimal control problem.