6G-DTFP: A Digital-Twin-Enabled Privacy-Preserving Federated User Prediction Framework for 6G Mobile Edge Computing
Cong Li, Lijing Zheng, Xinsheng Ji, Xingxing Liao, Zilong Wang, Guoqiang Mao · IEEE Communications Magazine · 2025
MEC is set to play a pivotal role in 6G, supporting the Internet of Everything. 3GPP has proposed user mobility analysis and prediction methods to manage vast user data. However, by bringing services closer to the network edge, MEC improves user access efficiency and exposes user predictions, including location data, to greater privacy risks and potential malicious attacks. Additionally, conducting experiments on large-scale user populations increases communication and training costs. To address these challenges, we propose the Digital-Twin-Enabled Privacy-Preserving Federated User Prediction Framework (6G-DTFP) for 6G MEC. This architecture incorporates a personalized model, enhances training efficiency, and strengthens privacy protection using differential privacy mechanisms. By leveraging Digital Twin technology, it maps real user entities to the virtual environment, improving insights into user characteristics and optimizing resource utilization. Experimental results show that this framework offers reliable user prediction and aligns with the sustainable development goals of 6G networks.