An H∞ filter based approach for battery SOC estimation with performance analysis

Abstract

State of charge(SOC) estimation is key to the battery management system used on electric vehicles. Recent studies often focus on the extended Kalman filter method for its theoretically optimal estimation property, but its effectiveness heavily depends on the priori information of noises and high precise battery models, therefore it is unable to deal with complex noises, and tend to fail on extreme environments. A new SOC estimation method using H∞ filter algorithm is presented to estimate SOC online. The algorithm is implemented on a computer, where its performance under different parameters is analysed, and a comparison with Kalman filter showed its robustness against colored noise.