New Framework Improves Jamming Detection in 5G Networks
/ 1 min read
🤖📶 New Federated Learning Framework Enhances Jamming Detection in 5G Networks. A novel two-stage federated learning framework has been developed to improve jamming detection in 5G femtocells, addressing the rising threat of complex jamming attacks that compromise network performance. The framework utilizes the Federated Averaging (FedAVG) algorithm for unsupervised training of a Convolutional Autoencoder (CAE), followed by a supervised classification stage using a fully connected network (FCN) trained with the Federated Proximal (FedProx) algorithm. Experimental results demonstrate high performance metrics, including a precision of 0.94 and an accuracy of 0.92, while ensuring data privacy and minimizing communication rounds to 30, making it a promising solution for secure 5G network operations.
