Online Health Monitoring and Anomaly Detection in Li-Ion Batteries via Trustworthy AI

This project aims to establish a new online battery health monitoring framework to track normal battery capacity loss and detect abnormal and safety-critical conditions. This project is significant because it will address specific challenges of sparse labelling, data heterogeneity/drift, AI model interpretability and privacy concerns, and prevent various disasters, e.g., destructive fire. The anticipated outcomes can significantly improve the trustworthiness of the AI system for battery health monitoring and help cultivate a local battery health monitoring industry timely. They will largely mitigate social concerns about battery safety and substantially broaden their applications in reducing greenhouse emissions to track climate change.