Table 2 The comparison of the literature review.
[Refs.] | Main algorithm used | Input parameters | Dataset used | Performance metrics |
|---|---|---|---|---|
Self-supervised transformer model (Deep learning) | Voltage, current, temperature | LG 18650 cell data from various drive cycles and temps | RMSE, MAE | |
Machine learning (SVM, GPR, Ensemble, NN) | Current, voltage, capacity, temperature | Panasonic 18650PF dataset | Not mentioned | |
GRU | Voltage, current, temperature | Panasonic 18650PF Dataset and Samsung 18650-20R Dataset | MAX, MAE | |
BGRU with TPE optimization | Not mentioned | Custom dataset from battery test bench | RMSE, MAE | |
LSTM + Bayesian optimization | Voltage, current, temperature | LG 18650 HG2 dataset with drive cycles | RMSE, MAE | |
ELM + GSA | Current, voltage, temperature | BJDST, US06 cycles | RMSE, MAE, MSE, MAPE | |
Improved ELM + SSA/SCA-SSA | Current, voltage, temperature | Custom dataset from energy storage device | MAE, MAPE | |
ELM + GSA | Current, voltage, temperature | BJDST, US06 cycles | RMSE, MSE, MAE, MAPE | |
Support vector regression | Not mentioned explicitly | Battery dataset with drive cycles | RMSE, Max error | |
Artificial neural network | Voltage, current, temperature, avg. voltage, avg. current | In-house BMS dataset | MSE, MAE, Max error | |
Proposed methodology | ANN, SVR and GPN with BO hyperparameter selection | Humidity, atmospheric temperature, battery V, I, T FET and motor temperature | Custom dataset from EV | RMSE, MSE and MAE |
