An improved H-infinity filter for SOC estimation of lithium-ion batteries based on fractional order model

A fractional order model is constructed to approximate the internal reactions of lithium-ion batteries. In the Hybrid Pulse Power Characteristic experiment, the parameter identification of the established model is achieved by particle swarm optimization algorithm with dynamic inertia weight. Based on the fractional order H-infinity filter, an improved H-infinity filter combined with the sliding mode observer is proposed. Through the verification of Urban Dynamometer Driving Schedule conditions, compared with the fractional H-infinity filter, the mean absolute error and the root mean square error of the state of charge (SOC) reduce from 0.88% and 1.12% to 0.76 % and 0.86%, respectively, and the SOC error of the proposed algorithm is less than 2%.