Prediction of Remaining Useful Life of Lithium-ion Battery Based on UKF

In the electric vehicle industry, lithium-ion batteries play an important role as a source of energy supply. However, there is a problem accurately predict its remaining useful life (RUL) of, which need to be solved urgently. This article describes the non-linear empirical models degradation capacity of the lithium ion battery. Under the framework of Kalman Filter (KF), an Unscented Kalman Filter (UKF) algorithm that can solve nonlinear problems effectively is selected. The model and algorithm are validated by simulation using NASA's open datasets, and then compared with existing Extended Kalman Filter (EKF) methods. The results demonstrated that UKF's prediction accuracy for RUL is improved by 5.2% compared to EKF, and the study discovered that UKF's prediction accuracy will not decrease with battery aging.