Design of a Low-Cost Artificial Intelligence Based Battery Management System

This research presents the design and implementation of a low-cost Artificial Intelligence-Based Battery Management System (AI-BMS) for lithium-ion batteries used in portable devices, solar power systems, and small electric vehicles. The study addresses thelimitations of existing systems-traditional thres hold based BMS that offer only reactive protection, and commercial smart BMS that are prohibitivelyexpensive and powerhungry.Using real degradation data from the NASA Prognostics Data Repository,an XGBoost machine learning model was developed to predict the State of Health (SoH) anddetect early thermal runaway precursors through voltage, current, and temperature trends. The trained model was deployed on an ESP32 microcontroller, integrated with low-cost sensors (INA3221 and ADS1115) and a 128×64 LCD for live system feedback.The AI-BMS achieved ±1.87% SoH accuracy, 100% safety response in fault simulations, and an average response time of 0.82 seconds, all at a total cost of approximately ₦22,450 ($18). Compared to conventional threshold-only protection and mid-range commercial BMS units, the proposed system offers proactive fault detection, predictive analytics, and real-time monitoring at a fraction of the cost and power consumption (35–48 mA). This study demonstrates that affordable, intelligent, and locally assembled BMS solutions can significantly enhance battery safety, extend lifespan, and democratize access to advanced energy storage technologies in developing regions.