State-of-charge estimation for battery management system using optimized support vector machine for regression

State-of-charge (SOC) estimation is one of the most challengeable tasks for battery management system (BMS) in electric vehicles. Since the external factors (voltage, current, temperature, arrangement of the batteries, etc.) are complicated, the formula of SOC is difficult to deduce and the existent SOC estimation methods are not generally suitable for the same vehicle running in different road conditions. In this paper, we propose a new SOC estimation based on an optimized support vector machine for regression (SVR) with double search optimization process. Our developed method is tested by simulation experiments in the ADVISOR, with a comparison of the estimations based on artificial neural network (ANN). It is demonstrated that our method is simpler and more accurate than that based on ANN to deal with the SOC estimation task.