Comparative Study of EKF and UKF for SOC Estimation of Lithium-ion Batteries

With the expanding application of lithium-ion batteries, the state of charge (SOC) estimation based on the Kalman filter (KF) draws great attention. However, lots of KF's variants, such as extended Kalman filter (EKF) and unscented Kalman filter (UKF) are available. In order to find the suitable KF variant for SOC estimation, this paper studies the performance of EKF and UKF in three scenarios based on the two-order RC model by elaborative simulation. The accuracy and convergence of the two filters are compared. The impact of the model deviation on the estimated results is also tested to evaluate the robustness. The test results show that EKF performs much the same as UKF in its accuracy, rate of convergence as well as robustness to model errors in most cases.