State of Charge Estimation Methods for Electric Vehicles & Hybrid Electric Vehicles: Present Standing and Upcoming Outlook – A Review

Accurate and continuous State of Charge (SOC) estimation is essential for the reliability, performance, and safety of electric vehicle (EV), hybrid electric vehicle (HEV) and battery management systems (BMS). This key metric directly influences driving range, reduces range anxiety, and is essential for improving battery performance and increasing its life. SOC can’t be measured directly but must be estimated from parameters like voltage, current, and temperature. This presents a significant challenge because of battery's non-linear and time-varying properties, changes caused by aging, and environmental factors like temperature and measurement noise. This review categorizes SOC estimation methodologies into direct measurement techniques, model-based, adaptive filter-based, and data-driven algorithms. While direct methods are simple, they often suffer from accumulated errors. Model-based and adaptive filter-based approaches offer higher precision and robustness, making them reliable for real-time EV applications. Data-driven methods, particularly deep learning, are increasingly prominent because of their superior precision and ability to handle complex non-linear battery behaviour, especially with large datasets. The current consensus points towards hybrid techniques, which combine the strengths of multiple methods, as the most promising and reliable solution for dynamic EV environments. Keywords: Artificial Intelligence, Battery Health Management, Battery Management System, Energy Storage System, Electric vehicles, Hybrid Energy Storage System, Hybrid Electric Vehicles, Lithium-ion, Machine Learning, State of Charge.