Physical Layer Authentication with Cooperative Wireless Communications and Machine Learning

Opposed to classical authentication protocols following a computational security paradigm based on secret credentials and cryptographic primitives, physical layer authentication aims at distinguishing users without shared secrets, by leveraging the natural randomness and uniqueness of transmission channels. We consider the special setting of cooperative wireless communications, in which some relay nodes are located between a supplicant and an authenticator, and we assess the performance of physical layer authentication approaches based on both statistical and machine learning techniques. We show that the presence of relay nodes enabling cooperative communications may improve the performance of physical layer authentication.