Reliable resource allocation with RF fingerprinting authentication in secure IoT networks

The unprecedented growth of the Internet of Things (IoT) has led to a huge amount of wireless resource consumption in a network. Due to limited wireless resources, a network can only guarantee the quality of service (QoS) of authenticated users rather than that of all users. By acknowledging this limitation, we realise that user authentication would be a big issue in IoT networks. Although traditional authentication methods can enhance network security to a certain extent, their vulnerability to malicious attacks and the relevant complicated computations restrict IoT deployments. In this paper, a radio frequency (RF) fingerprinting based authentication scheme is proposed under the architecture of convolutional neural network (CNN). It can effectively prevent unauthenticated users from consuming valuable wireless resources and significantly improve QoS performance for legitimate users. By solving an NP-hard optimization problem with the objective of minimizing efficient energy density, we demonstrate an approximate optimal resource allocation scheme in consideration of an RF-fingerprinting based authentication process. The analytic results show that our proposed scheme can dramatically reduce the efficient energy density compared with traditional cryptography based authentication schemes.