Transfer Learning-Based Physical Layer Authentication for Wireless Networks Handover

In dynamic access networks, terminals face frequent network handover which requires authentication handover to ensure the security of network access. Physical layer authentication based on machine learning will cause large computation overhead and communication delay for network. Therefore, this paper designs a fast authentication handover mechanism based on transfer learning, which makes full use of the source network authentication model and adopts transfer learning algorithm to transfer the previously trained model to target network. Target network continues to train on the basis of the model trained by source network, which simplifies the process of retraining authentication model during network handover greatly.