Real-time Reinforcement Learning Optimized Energy Management for a 48V Mild Hybrid Electric Vehicle 2019-01-1208
Energy management of hybrid vehicle has been a widely researched area. Strategies like dynamic programming (DP), equivalent consumption minimization strategy (ECMS), Pontryagin’s minimum principle (PMP) are well analyzed in literatures. However, the adaptive optimization work is still lacking, especially for reinforcement learning (RL). In this paper, Q-learning, as one of the model-free reinforcement learning method, is implemented in a 48V mild hybrid electric vehicle (HEV) framework to optimize the fuel economy. Different from other RL work in HEV, this paper considers not only battery state-of-charge (SOC), but also vehicle speed and vehicle torque demand as the Q-learning states. In the cost function definition, the fuel consumption contains engine fuel consumption and equivalent battery fuel consumption, which shares the idea with ECMS. The Q-value table is trained over one driving cycle multiple times. During the training process, the exploration and exploitation is discussed. In addition, the Q-learning performance is evaluated at different number training cycles to prove the realism of the strategy. After the training, the Q-learning strategy is compared with ECMS at different driving cycles. The results show that the fuel economy gets significant improvement by the Q-learning strategy.
Bin Xu, farzam malmir, Dhruvang Rathod, Zoran Filipi