Data-driven Energy Consumption Prediction of Electric Bus on a Given Route

The energy consumption prediction of electric vehicle (EV) is crucial for driving range estimation, eco-routing planning, charging station deployment and so on. This paper proposes a data-driven energy consumption prediction model for electric bus on a given route using support vector regression (SVR) method. First, the driving data of an electric bus over a span of 1 year at a particular given route are collected, and the invalid trips are eliminated to avoid the inefficient training. Second, an EV energy consumption prediction model based on the SVR method is proposed considering the impacts of driving speeds, environmental temperature and driving conditions. Finally, simulations results show that the relative mean errors of the proposed prediction method were 7.4% and 10.0% for training and test dataset, respectively.