AI POWERED HYBRID MACHINE LEARNING AND RNN BASED PREDICTIVE BATTERY MANAGEMENT SYSTEM FOR ELECTRIC VEHICLES

With the growing shift toward electric mobility, the need has increasingly become critical for reliable and smart Battery Management Systems (BMS). This work details development of an AI driven hybrid Predictive Battery Management System for electric vehicles delivering advanced diagnostics plus safety controls using embedded technologies. The proposed system integrates a suite of sensors along with an embedded platform to continuously monitor battery voltage, current, temperature, state of charge, state of health, and physical condition. Data transmits to a dashboard using the cloud simultaneously plus processes on the device locally to respond promptly for remote visualization. Historical battery data is used in order to train machine learning models including Autoencoders, Random Forest classifiers, as well as Long Short-Term Memory (LSTM) networks. These models are fine tuned to identify abnormal behavior and predict the deterioration. This allows for predictive maintenance. Layered AI is considered as it is a fusion of time series analysis and classification methods, which can lead to enhanced fault prediction. Furthermore, incorporating certain edge computing capabilities means that the solution can take its own independent safety measures in urgent situations instead of relying on cloud infrastructure. This work presents prototype that work with the Raspberry Pi hardware and custom sensors. It proves the enhanced operational safety, minimize sudden accidents and prolong battery life.