eXogenous Kalman Filter for Lithium-Ion Batteries State-of-Charge Estimation in Electric Vehicles

TL;DR

This paper introduces the eXogenous Kalman filter (XKF), a two-stage nonlinear estimator for accurately estimating the state-of-charge in lithium-ion batteries within electric vehicles, demonstrating improved convergence over traditional methods.

Contribution

The paper proposes a novel two-stage nonlinear estimator combining a nonlinear observer and a linearized Kalman filter for better SOC estimation in EV batteries.

Findings

XKF achieves faster convergence than EKF and UKF.

The method performs well under different temperature conditions.

Experimental validation on LiFePO4 batteries confirms effectiveness.

Abstract

This paper presents a novel framework for state-of-charge estimation of rechargeable batteries in electric vehicles using a two-stage nonlinear estimator called the eXogenous Kalman filter (XKF). The nonlinear estimator consists of a cascade of nonlinear observer (NLO) and linearized Kalman filter (LKF). The NLO is used to produce a globally convergent auxiliary state estimate that is used to generate a linearized model in the time-varying Kalman filter algorithm. To demonstrate the proposed approach, we present a model of a lithium-ion battery from an equivalent circuit model (ECM). The model has linear process equations and a nonlinear output voltage equation. The method is tested using experimental data of a lithium iron phosphate (LiFePO$_4$) battery under dynamic stress test (DST) and federal urban driving schedule (FUDS). Effect on different ambient temperatures is also discussed. Compared with EKF and UKF, our proposed XKF achieve faster convergence rate, which can be attributed to the use of the NLO.