Comparative Analysis of Extended and Unscented Kalman Filters for SoC Estimation in Lithium-Ion Batteries: A Model-Based Approach

Accurate state of charge (SoC) estimation is critical for lithium-ion batteries, especially under dynamic load conditions. This paper compares the performance of the extended Kalman filter (EKF) and the unscented Kalman filter (UKF) for SoC estimation using a MATLAB/Simulink-based model. To simulate battery dynamics, a second-order equivalent circuit model is used. Both filters are implemented under identical initial conditions and noise characteristics. Evaluation metrics include estimation error, convergence rate, and numerical stability. Simulation results show that while EKF and UKF achieve comparable accuracy in steady-state operation, the UKF demonstrates improved performance during transient conditions. Despite the theoretical advantages of UKF in handling nonlinearities, its practical benefits are most evident during dynamic phases. These results indicate that filter selection may consider transient behavior and computational constraints depending on the requirements.