Privacy-Preserving Method for EV Charge Location Prediction
/ 1 min read
🔋✨ Innovative Approach Enhances Privacy in Electric Vehicle Charge Location Prediction. As electric vehicles (EVs) are expected to dominate global vehicle sales by 2050, concerns about data privacy in energy demand predictions have emerged. Researchers developed a Federated Learning Transformer Network (FLTN) that allows EVs to predict their next charge location while safeguarding sensitive mobility data. This system enables each EV to train a model that shares only weights, not raw data, with a community-based Distributed Energy Resource Management System (DERMS). The FLTN achieved up to 92% accuracy, balancing privacy and performance, compared to a centralized model that reached 98% accuracy without privacy measures. This approach effectively mitigates data leakage risks while forecasting energy demands over extended periods.
