Predicting the State of Health of Lithium Battery Based on Improved LSTM-RNN

With the increasingly widespread use of lithium-ion batteries in electric vehicles, much research has been done on the safe and reliable operation of the battery management system (BMS). Accurate estimation of SOH, which is one of the key functions of BMS, is of great importance for the safe operation of electric vehicles. we propose a time series model to predict the SOH of Li-ion batteries, which is based on a modified long short-term memory recurrent neural network architecture (LSTM-RNN), and we also use an improved particle filtering algorithm to optimize the SOH prediction. The experimental results show that the improved LSTM-RNN scheme has improved prediction results with an average absolute error of 1.4% and an RMSE of 1.8%.