Secure Anomaly Detection in EV Batteries via Federated Learning

As the electric vehicle industry continues to grow in popularity in the fight against greenhouse gas emissions, so does the demand for accurate and timely battery health monitoring systems. However, a much larger challenge is maintaining privacy while transmitting sensitive battery information during the anomaly detection process. We present a privacy-preserving federated learning framework that utilizes Cheon-Kim-Kim-Song (CKKS) homomorphic encryption, enabling secure anomaly detection in electric vehicle batteries. The approach reduces privacy risks by enabling collaborative model training without the need to transmit raw sensor data. We evaluated the framework on a real-world electric vehicle battery dataset and found that it performs better than traditional centralized methods with a minimal computational overhead. The proposed framework presents a practical and secure approach to anomaly detection in electric vehicle batteries, which is a central challenge to the widespread adoption of electric vehicles.