EXPLAINABLE DATA DRIVEN DIGITAL TWINS FOR PREDICTING BATTERY STATES IN ELECTRIC VEHICLES

As the automotive sector acceleratestowards electric vehicles (EVs), predicting batterystates accurately is vital for maximizing performance,safety, and lifespan. This project presents a novelapproach that utilizes Explainable Data-Driven DigitalTwins to forecast battery states in electric vehicles(EVs). It incorporates various advanced machinelearning algorithms, including DNN, LSTM, CNN,SVR, SVM, FNN, RBF, RF, and XGBoost. The keyobjective is to enhance the accuracy of predictingcritical battery metrics like SOC and SOH underdiverse operating conditions. Additionally, the projectapplies explainable AI to uncover factors that impactbattery performance. By harnessing the strengths ofvarious algorithms, the digital twin model showsimproved prediction accuracy and resilience comparedto traditional methods. This research advancesintelligent, adaptive battery management systems,paving the way for the future of electric transportation.