AI-ENABLED BATTERY MANAGEMENT SYSTEM FOR REAL-TIME SOC AND SOH ESTIMATION IN ELECTRIC VEHICLES

Abstract

This research paper introduces an AI-driven Battery Management System (BMS) to estimate State of Charge (SOC) and State of Health (SOH) of lithium-ion batteries in real-time to be used in electric vehicles. The suggested framework will combine embedded sensing with ESP32 microcontroller, hybrid estimation with Coulomb counting and Extended Kalman Filter (EKF), and machine learning models such as Random Forest, XGBoost, and LSTM. Filtering, normalization, and feature engineering are used to preprocess real-time battery data which includes voltage, current, and temperature to improve the performance of models. The experimentally verified hardware prototype of the system is a multi-cell lithium-ion battery pack, sensing modules, relay-based protection, and cloud-based monitoring based on MQTT. The outcomes indicate that the estimation is very accurate, and the quality of R2 is over 0.98, and the error in SOC is 1-2%. The system can be operated with good performance in different temperature conditions and allows good monitoring and fault detection and protection, which makes it appropriate in the practical management of battery in electric vehicles.