GOMES, J. P. M.; http://lattes.cnpq.br/0873370884714415; GOMES, João Pedro Melquiades.
Resumo:
Physical layer security is a relatively unexplored area in the context of 5G networks. IoT devices are susceptible to counterfeiting with the intention of inĄltrating a network, acquiring sensitive information while posing as authenticated devices. This work involves the implementation of device identiĄcation systems based on the manufacturing characteristics of the antenna arrays that transmit signals in the network, utilizing various deep learning methods to address this issue. As a result, the models achieved a speciĄcity of 100% after training, effectively detecting all invading devices. This demonstrates the feasibility of leveraging artiĄcial intelligence algorithms to safeguard high-speed transmission networks emerging with 5G, providing protection that is independent of transmitted data and relies on difficult-to-duplicate characteristics.