Capacity Estimation for Lithium-ion Batteries Using Adaptive Filter on Affine Space

In this paper, we propose methods to estimate the full charge capacity (FCC) of a battery based on adaptive filters. The FCC is estimated as a ratio of the accumulated chargecurrent to the state of charge (SoC) of the battery, which is estimated by an extended Kalman filter. We consider bias errors on the estimated SoC caused by the error of typical value of FCC, which is assumed in the SoC estimation. We also consider the current sensor offset, which causes unboundedness of variables in the FCC estimation. We compose the adaptive filters on an affine space to avoid the unboundedness, which is undesirable for an implementation in embedded systems.