Enhanced lithium-ion battery state of charge estimation in electric vehicles using extended Kalman filter and deep neural network

In an electric vehicle, it is crucial to accurately determine the remaining energy in the battery pack, commonly referred to as the state of charge. Obtaining this information through direct measurement in such applications is often challenging. To address this issue, an algorithm that combines an extended Kalman filter and deep neural networks was developed using Matlab Simulink. The results demonstrate that the proposed strategy achieves the highest possible accuracy in estimating the state of charge. The output of the model has consistently been more accurate, with a predicting error for the state of charge that is less than 1.59%. This demonstrates the effectiveness and efficiency of this method. This approach is already applicable in practical applications.