An Improved Robust H ∞ Observer Framework for State-of-Charge Estimation of Lithium-Ion Batteries in Electric Vehicles

Accurate state-of-charge (SOC) estimation is a critical requirement for reliable battery management systems (BMS) in electric vehicle (EV) applications, especially for lithium-ion batteries (LIBs) operating under dynamic conditions. However, variations in load demand, modeling uncertainties, and external disturbances can significantly degrade the performance of conventional SOC estimation techniques. To address these challenges, this paper proposes a robust SOC estimation framework based on anH∞observer, specifically designed to attenuate worst-case disturbances and enhance estimation robustness. The effectiveness of the proposed method is validated using experimentally measured current profiles from standard driving cycles, including the Urban Dynamometer Driving Schedule (UDDS), Supplemental Federal Test Procedure (US06), Los Angeles 1992 (LA92), and Highway Fuel Economy Test (HWFET). Initially, the proposed observer is compared with Kalman filter (KF)-based variants under identical modeling assumptions and operating conditions, achieving an overall root mean square error (RMSE) of 0.105%, which corresponds to an SOC estimation accuracy of nearly 99.9%. Furthermore, benchmarking against recent SOC estimation approaches reported in the literature demonstrates an average performance improvement of 82%. These results confirm the superior accuracy, robustness, and real-time applicability of the proposedH∞observer for practical EV lithium-ion BMS.