A Hybrid Kalman Filter and Machine Learning Framework for Data-Driven SoH Estimation of Lithium-Ion Batteries

Abstract

Accurate estimation of a battery's State of Health (SoH) is essential for optimizing battery management systems (BMS) and enhancing the lifespan and reliability of battery-operated devices. This paper introduces a novel approach that integrates various machine learning algorithms, including decision tree methods, linear regression, Gaussian Support Vector Machines (SVM), neural networks, and kernel methods, to predict SoH based on real-time data from voltage, current, and temperature sensors. By employing performance metrics such as root mean square error (RMSE) and R-squared (R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>), the study evaluates the effectiveness of these models in providing reliable SoH assessments. The findings emphasize the significance of advanced predictive modeling in battery management, enabling proactive maintenance and optimizing charge/discharge cycles. This work highlights the potential of machine learning to enhance battery management practices, particularly in applications such as electric vehicles and renewable energy storage.