An Optimization Framework for Active Physical-Layer Authentication

Abstract

This paper concerns the problem of the parameter optimization of an active Physical-Layer Authentication (PLA) scheme, which is crucial for minimizing the distortion on the base signal carrying the original message caused by embedding a tag. An inappropriate parameter may significantly lower the efficiency of the entire period of message transmission. In this paper, we propose an optimization framework for an active PLA scheme and further propose a new systematic metric, defined as Secure Authentication Efficiency (SAE). In the proposed framework, we minimize the aforementioned distortion by tuning three parameters, i.e., the Probability of Message Transmission (PMT), the Probability of Message Outage (PMO), and the Probability of Secure Authentication (PSA). The proposed optimization framework deepens the understanding of the correlation between the parameters of an active PLA scheme and the conditions of a wireless channel, and allows us to systematically optimize the parameters of the active PLA scheme. Then, we establish an objective function by maximizing the SAE with both PMT and PMO constraints. We implement our approach and conduct extensive performance comparisons. Our experimental results show that when the SNR at the receiver is more than 15 dB, our approach achieves an average SAE of greater than 80%, whereas the prior scheme without parameter optimization degenerates to zero SAE under some conditions, e.g., high SNR at adversary or high communication rate.