Energy management strategy for plug-in hybrid electric vehicles based on vehicle speed prediction and limited traffic information

—The adaptability of the energy management strategy (EMS) of a plug-in hybrid electric vehicle (PHEV) to actual road conditions determines the actual performance. This paper proposes a long short-term EMS (LS-EMS) that utilizes limited traffic information and speed prediction to constrain the state of charge (SOC) and optimize the powertrain. The strategy aims to improve the fuel economy of a power-split PHEV on real driving routes. In the long-time scale, the fuzzy controller converts limited connected information into SOC constraints for each sub-section of the entire trip. In the short-time scale, a speed predictor constructed with temporal convolutional networks, predicts vehicle speed from 5 to 15 s into the future. Combining the long short-term information, the model predictive control optimizes power allocation based on the look-ahead vehicle speed under segmental SOC constraints. Simulations are conducted on the collected commuter routes. The proposed method within a 10-s prediction period closely approximates the optimal result calculated by dynamic programming. The proposed LS-EMS plans the range of SOC by section on the commuter road and reduces the energy consumption by 12.0 % compared to the charge-depleting charge-sustaining rule strategy.