Physical Layer Authentication Based on Gaussian Mixture Model Under Unknown Number of Attackers

Message authentication based on wireless physical layer channel information has gained significant attention in recent years. In existing studies, there are several channel based authentication methods to deal with the single attacker scenario. However, in the real wireless environment, there may be several attackers and we do not know the exact number of the attackers. To solve the physical layer authentication problem in such a multi-attackers scenario, we propose a variational Bayesian algorithm based authentication scheme using Gaussian mixture model. We show that even without having a complete prior knowledge and the number of the attackers, our algorithm can identify the received messages to determine whether they are from the legitimate transmitter or the attackers. We experimentally demonstrate the performance of our proposed method and show that the variational Bayesian algorithm has a low miss detection rate.