Online Parameter Optimization of an Adaptive Extended Kalman Filter Using the Starfish Optimization Algorithm for Lithium-Ion Battery State-of-Charge Estimation

Abstract

: Accurate state-of-charge (SOC) estimation is an essential function of battery management systems (BMSs). Model-driven methods are widely used for SOC estimation because of their high estimation accuracy and moderate computational cost. However, the performance of the adaptive extended Kalman filter (AEKF) is highly dependent on the settings of the noise covariance parameters, which generally require case-specific tuning for different batteries and operating conditions. Accordingly, an starfish optimization algorithm (SFOA)-AEKF framework is developed to adaptively optimize the AEKF noise covariance parameters under dynamic operating conditions using the starfish optimization algorithm (SFOA). First, the battery model parameters are identified online using forgetting factor recursive least squares (FFRLS) and incorporated into the AEKF for SOC estimation. Subsequently, SFOA uses historical data from each optimization interval to update the filter noise covariance parameters, allowing the AEKF to adapt to current operating conditions. The weighting coefficient is also evaluated to determine the balance between SOC and terminal-voltage errors during optimization. The proposed method is evaluated under the US06 Driving Schedule (US06), Federal Urban Driving Schedule (FUDS), and Beijing Dynamic Stress Test (BJDST) driving cycles at 25 ○ C and 45 ○ C with both matched and mismatched initial SOC conditions and is compared with Extended Kalman Filter (EKF), AEKF, particle swarm optimization (PSO)-AEKF, genetic algorithm (GA)-AEKF, and grey wolf optimization (GWO)-AEKF. The results show that SFOA-AEKF improves SOC estimation accuracy, convergence, and robustness under different driving cycles, temperatures, and initial SOC conditions. The method is suitable for SOC estimation in dynamic applications, particularly in electric vehicle Battery Management Systems (BMSs).