Fractional Extended and Unscented Kalman Filtering for State of Charge Estimation of Lithium-Ion Batteries

Two state of charge estimation methods using fractional order extended and unscented Kalman filter and a nonlinear variable fractional order battery model are implemented. Both, battery model and Kalman filters are evaluated and compared using measurements of an actual lithium-ion polymer battery cell. The observability of the battery model and the influence of an initialization function on the estimation algorithms is investigated.