State of Charge Estimation of Lithium-ion Batteries Using Long Short Term Memory Network in Electric Vehicle Applications

Precise State-of-Charge (SoC) estimation is crucial for accurate electric vehicle range prediction and efficient battery charge management. This paper explores a computationally efficient machine learning approach to address SoC-estimation-related issues. A Long Short-Term Memory (LSTM) network is utilized for predicting the SoC of lithiumion batteries, leveraging real-time data on temperature, voltage and current. The proposed LSTM model demonstrates remarkable performance, achieving a mean absolute error of 1.02%, as well as a 1.35% root mean squared error at 25°C, while ensuring low computational overhead, making it suitable for practical implementation in Battery Management Systems (BMSs). This work lays a solid foundation for future advancements in BMS design, offering significant improvements in SoC prediction accuracy, reliability and efficiency. The findings suggest that LSTM networks can effectively model the complex dynamics of battery behavior, even when having varied operational conditions.