ML Based Prediction Of Battery SOH & RUL

Abstract

This research focuses on battery health monitoring and life prediction for laptops and mobile devices using real-time data. The system is designed to analyze battery performance and predict the State of Health (SOH) and Remaining Useful Life (RUL) using Machine Learning techniques. The proposed system is deployed on an online platform and uses actual battery reports generated through system commands. The battery report is uploaded to a web-based dashboard, where data is extracted, cleaned, and processed for analysis. The system uses a Long Short-Term Memory (LSTM) model along with optimization techniques such as Particle Swarm Optimization (PSO) and attention mechanism to improve prediction accuracy. The output is displayed in an interactive dashboard with battery health status, degradation level, and graphical representation. The system currently supports laptop battery analysis and is being extended for mobile battery monitoring through USB-based integration. Overall, the system provides a practical, scalable, and user-friendly solution for battery health prediction