Data-driven remaining driving range estimation and analysis framework for electric vehicles under real-world conditions ¹

TY - JOUR

T1 - Data-driven remaining driving range estimation and analysis framework for electric vehicles under real-world conditions 1

AU - Zhou, Litao

AU - Wang, Zhenpo

AU - Mao, Zhiyu

AU - Wang, Qiushi

AU - Zhang, Dayu

AU - Zhang, Zhaosheng

AU - Chen, Zhongwei

N1 - Publisher Copyright: © 2025

PY - 2026/1/1

Y1 - 2026/1/1

N2 - Range anxiety is one of the major issues impeding the widespread adoption of electric vehicles. Accurate estimation of the remaining driving range (RDR) can effectively address this problem. However, in real-world operations, the coupling effects of various factors such as driving behavior, ambient temperature, and battery aging present significant challenges to accurate RDR estimation. In this study, a novel data-driven framework is proposed to achieve real-time online estimation and analysis of the remaining driving range for real-world vehicles. Firstly, methods for calculating energy consumption and state of health (SOH) based on real vehicle operation data are proposed. Then, stepwise estimation of the energy consumption rate and RDR is conducted using random forest regression methods. Finally, a multidimensional analysis of driving range capability is achieved through the estimation model. The proposed framework was validated on passenger vehicle and bus fleets in different cities over a three-year period. The results show that the prediction accuracy for the driving range reached a mean relative error of 5.39 % and 5.45 %, respectively. By adjusting driving behavior, the driving range can be improved by over 30 % for passenger vehicles and over 10 % for buses.

AB - Range anxiety is one of the major issues impeding the widespread adoption of electric vehicles. Accurate estimation of the remaining driving range (RDR) can effectively address this problem. However, in real-world operations, the coupling effects of various factors such as driving behavior, ambient temperature, and battery aging present significant challenges to accurate RDR estimation. In this study, a novel data-driven framework is proposed to achieve real-time online estimation and analysis of the remaining driving range for real-world vehicles. Firstly, methods for calculating energy consumption and state of health (SOH) based on real vehicle operation data are proposed. Then, stepwise estimation of the energy consumption rate and RDR is conducted using random forest regression methods. Finally, a multidimensional analysis of driving range capability is achieved through the estimation model. The proposed framework was validated on passenger vehicle and bus fleets in different cities over a three-year period. The results show that the prediction accuracy for the driving range reached a mean relative error of 5.39 % and 5.45 %, respectively. By adjusting driving behavior, the driving range can be improved by over 30 % for passenger vehicles and over 10 % for buses.

KW - Lithium-ion batteries

KW - electric vehicles

KW - machine learning

KW - remaining driving range estimation

UR - https://www.scopus.com/pages/publications/105020919652

U2 - 10.1016/j.apenergy.2025.126881

DO - 10.1016/j.apenergy.2025.126881

M3 - Article

AN - SCOPUS:105020919652

SN - 0306-2619

VL - 402

JO - Applied Energy

JF - Applied Energy

M1 - 126881

ER -