Intelligent energy consumption prediction for battery electric vehicles: A hybrid approach integrating driving behavior and environmental factors

The precise prediction of energy usage in Battery Electric Vehicles (BEVs) effectively mitigates drivers' concerns over "mileage anxiety". However, the conventional approach to predicting energy consumption, which relies solely on historical data and a single model, exhibits significant limitations in terms of accuracy and applicability. These limitations are particularly evident in scenarios lacking traffic information, where uncertainty about velocity and driving patterns can result in suboptimal predictions. As a result, a hybrid method based on driving style and route information recognition is proposed in this paper to accurately predict future energy consumption. This method relies on multi-source information and achieves its objective through a driving cycle prediction and residual fitting model. Simulation results indicate that the framework exhibits acceptable predictive performance in urban, motorway, and suburban settings, with Terminal Relative Errors (TRE) of 5.40%, 5.60%, and 4.26%, respectively.