SoC estimation based on adaptive EKF with colored noise

The paper utilizes a novel battery model based on the electrical features of LiFePO4 battery, because Kalman filter algorithm(KF) is largely dependent on system model. Measurements of battery state are easily disturbed by colored noise which is high relevance in working condition, and the paper studies that the system noise satisfy one-order AR model. The paper proposes an adaptive extended Kalman filter with colored noise(CN-AEKF). The simulation proves that the algorithm can accurately estimate SoC and possess algorithm stability. At the same time, the algorithm solves the sensitivity of initial covariance value of system noise, improving the adaptivity of estimating SoC.