URLLC Physical Layer Authentication based on non-linear Supervised Learning

In this work, non-linear supervised learning classifiers are proposed for Physical Layer Authentication (PLA). Their performance is evaluated using a Universal Software Radio Peripheral (USRP) based testbed under randomly distributed attacks and burst attacks within a mobile Ultra Reliable Low Latency Communication (URLLC) campus network scenario. It is shown, that in specific cases, non-linear classifiers can achieve promising results in terms of authentication accuracy and Receiver Operating Characteristics (ROCs) curve performance.