Estimation State of Charge for Lithium Battery Based on Adaptive Unscented Kalman Filter

This paper mainly studies the estimation problem of lithium battery SOC with time-varying noises in battery systems. First, a state space model of the lithium battery is established, and the model parameters of the battery system are obtained through parameter identification methods. Subsequently, an adaptive noises-resistant noises estimator for time-varying noises was designed, and an adaptive unscented Kalman filter (AUKF) with high filtering accuracy is proposed to obtain a more accurate SOC. Finally, the proposed algorithm is validated to be effective and highly accurate through lithium-ion battery experiment.