This project introduces a machine learning-based framework for accurately estimating the Remaining Useful Life (RUL) of electric vehicle (EV) batteries and implementing a fair pricing model in battery swapping stations. The existing system leverages XGBoost Regression to predict RUL based on operational parameters. To enhance prediction performance and system robustness, the proposed approach integrates multiple regression models including Random Forest Regressor, K-Nearest Neighbors (KNN), Stacking Regressor, and Voting Regressor. These models are trained on real-world battery features such as discharge time, cycle index, voltage thresholds, and charge duration. Ensemble methods such as stacking and voting are employed to combine the strengths of individual models, improving overall accuracy. The predicted RUL is used to dynamically calculate fair pricing, ensuring cost-effectiveness and reliability for EV users. This system supports predictive maintenance and promotes the sustainability of EV infrastructure by enabling data-driven battery lifecycle management in real time.

Keywords: Remaining Useful Life (RUL), Electric Vehicles (EV), Battery Swapping Stations, XGBoost, Random Forest, KNN, Stacking, Voting Regressor, Fair Pricing, Predictive Maintenance, Real-Time Estimation, Ensemble Learning.