Prediction of State of Charge in Electric Buses Using Supervised Machine Learning Techniques

The increasing adoption of battery-electric buses (BEBs) necessitates effective methods for managing their energy use, particularly in urban areas. This paper aims to accurately predict the State of Charge (SOC) of BEBs in Guangzhou, China, an essential factor for efficient energy management. We employed supervised machine learning techniques, specifically Random Forest and eXtreme Gradient Boosting (XGBoost), to develop predictive models for SOC. The study involved selecting eleven key features based on their inter-correlations to construct these models. The performance of the models was evaluated using the root mean square error (RMSE) metric. Results indicated an RMSE of 2.31 for Random Forest and 8.59 for XGBoost, demonstrating the effectiveness of these methods in predicting SOC with considerable accuracy. This research contributes to optimizing the operation of electric buses by providing reliable tools for SOC estimation, crucial for planning and managing urban public transportation systems.