Suggested Citation:
Li, Haoyu, Xinqi Xie, Xinyang Zhang, Andrew Burke, Jingyuan Zhao (2025)

Battery State Estimation for Electric Vehicles: Translating AI Innovations Into Real-World Solutions

Electrification of transportation is a crucial strategy to mitigate climate change and reduce air pollution. Battery electric vehicles (BEV) are central to this initiative, significantly reducing transport emissions. Yet, the optimal performance of BEV relies heavily on precise battery performance, particularly with respect to capacity degradation (state of health, SOH) and safety risks (state of safety, SOS). These challenges are critical as capacity degradation can impair vehicle performance and safety risks such as thermal runaway may have drastic consequences. The predictive modeling of battery life is hindered by factors such as inconsistencies in materials, varying manufacturing processes, changing operational conditions, and the diversity of data quality. To address these challenges, some cloud-based, artificial intelligence (AI)-enhanced framework that integrates longitudinal electronic health records with real-world operational data provides a robust solution. These advanced digital platform enables continuous and dynamic assessment and prediction of battery performance. In this review, we outline the current challenges, emerging techniques, and future directions within a unified framework designed to promote intelligent, interconnected battery management systems (BMS). These developments are essential for improving the reliability and efficiency of BEV, thereby facilitating the global transition toward sustainable transportation.