An energy management strategy (EMS) is important for hybrid electric vehicles (HEVs) since it plays a decisive role on the performance of the vehicle. However, the variation of future driving conditions deeply influences the effectiveness of the EMS. Most existing EMS methods simply follow predefined rules that are not adaptive to different driving conditions online. Therefore, it is useful that the EMS can learn from the environment or driving cycle. In this paper, a deep reinforcement learning (DRL)-based EMS is designed such that it can learn to select actions directly from the states without any prediction or predefined rules. Furthermore, a DRL-based online learning architecture is presented. It is significant for applying the DRL algorithm in HEV energy management under different driving conditions. Simulation experiments have been conducted using MATLAB and Advanced Vehicle Simulator (ADVISOR) co-simulation. Experimental results validate the effectiveness of the DRL-based EMS compared with the rule-based EMS in terms of fuel economy. The online learning architecture is also proved to be effective. The proposed method ensures the optimality, as well as real-time applicability, in HEVs.

Energy management strategies (EMSs) in hybrid electric vehicles (HEVs) are highly related to the fuel economy and emission performances. However, EMS constitutes a challenging problem due to the complex structure of a HEV and the unknown or partially known driving cycles. To meet this problem, this paper adopts a stochastic dynamic programming (SDP) method for the EMS of a specially designed vehicle, a pre-transmission single-shaft torque-coupling parallel HEV. In this parallel HEV, the auto clutch output is connected to the transmission input through an electric motor, which benefits an efficient motor assist operation. In this EMS, demanded torque of driver is modeled as a one-state Markov process to represent the uncertainty of future driving situations. The obtained EMS has been evaluated with ADVISOR2002 over two standard government drive cycles and a self-defined one, and compared with a dynamic programming (DP) one and a rule-based one. Simulation results have shown the real-time performance of the proposed approach, and potential vehicle performance improvement relative to the rule-based one.