Review of intelligent energy management techniques for hybrid electric vehicles

This paper presents a comprehensive review of energy management systems for hybrid electric vehicles with a focus on rule-based and reinforcement learning-based techniques. The authors strongly believe that the future of energy management systems involves a combination of both types of techniques discussed in this paper. The paper is complete in its subject as it begins with the basic architectures of hybrid electric vehicles followed by energy storage mechanisms in the hybrid electric vehicles leading into the discussion on energy management. The energy management discussion has been segregated into two types of approaches, i.e., online approaches and offline approaches where the offline approaches mostly lead to rule-based energy management policy whereas the focus of the online approaches discussed in the paper is on reinforcement learning-based methods. Our study has revealed some important research gaps in the existing energy management strategies. Specifically, there is a substantial lack of formal modeling regarding the uncertainty involved in the problem. Also, there is a lack of literature focusing on the utilization of multi-layer control approaches. We have pointed out modular and hierarchical control approaches for future energy management systems. Finally, the implementation issues for the energy management systems have been highlighted which include the instrumentation-related problems and consideration of user behavior as well as weather conditions.