DefinePK

Abstract

This research presents a Battery Monitoring and Notification System that leverages machine learning to predict the Remaining Useful Life (RUL) of vehicle batteries. The system analyzes key indicators such as charging cycles, voltage, temperature, discharge patterns, and battery retention, in combination with real-time sensor inputs and historical performance data. A Decision Tree Regressor is employed to deliver accurate RUL predictions, with model performance evaluated using Mean Absolute Error (MAE), Mean Squared Error (MSE), and R² Score. The system features automated email notifications for critical battery conditions, including low voltage, overcharging, and performance degradation. Continuous voltage monitoring enables early detection of potential issues, supporting proactive maintenance. Data is efficiently collected and stored in an SQLite database, while a user-friendly Flask-based dashboard offers visual insights into battery trends, charging history, and predictive analytics. This integrated approach enhances maintenance planning, reduces operational downtime, and improves safety by enabling timely, data-driven decision-making.