An energy management strategy for hybrid electric bus based on reinforcement learning

Hybrid electric buses become more and more popular in cities because of higher fuel efficiency and less emission pollution. The power split between internal-combustion engine and electric motor, known as energy management strategy (EMS), is an important issue in hybrid electric vehicles, which has a significant impact on the overall efficiency. In this paper, an energy management strategy for buses is proposed based on reinforcement learning, utilizing property that the bus runs on the same route again and again. With the self-learning EMS implemented, city buses can adapt to the driving condition automatically after some driving cycles. The benefits of the proposed strategy are shown by a simulation study using Advanced Vehicle Simulator (ADVISOR) in Matlab. The results suggest the proposed method achieves both better fuel economy and less emissions.