Battery management systems for electric vehicle: Technologies, challenges, and future research directions

Abstract

The rapid development of Electric Vehicle (EV) has increased the demand for battery management systems (BMSs) capable of ensuring safe, reliable, and efficient operation under dynamic conditions. This review critically examines the evolution of EV-BMS technologies from conventional monitoring and protection to AI-enabled, connected, and application-oriented systems. Unlike reviews that primarily address individual BMS functions or specific battery technologies, this work integrates battery characteristics and degradation with state estimation, thermal management, cell balancing, fault diagnosis, safety, and emerging BMS architectures, while emphasizing their interactions and practical deployment constraints. Recent developments in machine learning, deep learning, reinforcement learning, physics-informed methods, cloud-assisted and digital-twin-based BMS, wireless communication, cybersecurity, and emerging battery chemistries are comparatively examined in terms of computational burden, data requirements, safety, scalability, and real-time applicability. The review further identifies system-level challenges related to cross-chemistry transferability, uncertainty and sensor faults, AI validation, communication reliability, cloud-edge coordination, and BMS adaptation for emerging and second-life batteries. These findings highlight the need to move beyond isolated algorithmic improvements toward validated, interoperable, and application-aware BMS architectures for future EVs.