Data-driven estimation of driving distance for battery electric bus with multisource real-world data

Battery electric buses (BEBs) have been regarded as an environmentally friendly mode of public transportation to address the congestion and pollution problems. However, due to the limited driving ranges of BEBs, drivers frequently experience range anxiety, which has been identified as a primary obstacle to the promotion of BEBs. Accurately estimating driving distance can effectively mitigate the negative impacts of this concern. To this end, this study utilizes multi-source data collected from real-world situation to investigate the estimation of driving distance of BEBs. Firstly, based on the realworld data related to BEB operation and weather, three tree models are adopted to establish the models for driving distance estimation, including decision tree (DT), random forest (RF) and gradient boosting decision tree (GBDT). The models considers twelve impacting factors. Secondly, the performance of the tree models in estimation is assessed, revealing that they demonstrate strong capabilities in predicting driving distance, with GBDT achieving the highest accuracy among them. Lastly, the relative significance of the factors influencing the driving distance of BEBs is examined using the proposed models, followed by a discussion of valuable recommendations.