Advanced State of Charge Estimation for Electric Vehicle Batteries Using Gradient Boosting and Random Forest Models

The growing interest in artificial intelligence (AI) and machine learning (ML) has increased the research and development of new methods to estimate the state of charge (SoC) of batteries in electrified vehicles. The accurate estimation of the state of charge (SoC) is essential to ensure the reliability and autonomy of electric vehicle (EV) batteries. In this article, a Random Forest algorithm and a Gradient Boosting algorithm have been used to estimate the SoC from experimental data obtained in the laboratory on a lithium-ion battery using a lithium-ion battery cell. This article presents a comparative study between the SoC measured experimentally and the SoC estimated considering the effect of temperature and various current profiles, so the battery can be faced with the diverse situations possible. Both models aim to demonstrate the ability of artificial intelligence to estimate the SoC. The performance of the Random Forest model appears to be excellent on the validation data, with a mean absolute error (MAE) of less than 0.3% and a coefficient of determination of 0.999. In contrast, the Gradient Boosting model gives a mean absolute error (MAE) value of 1% and a coefficient of determination of 0.997. The results demonstrate that the Random Forest model is the most accurate and reliable.