Divergence-Minimizing Attack Against Challenge-Response Authentication with IRSs
Laura Crosara, Anna V. Guglielmi, Nicola Laurenti, Stefano Tomasin · 2024
We propose a new attack against challenge response physical layer authentication (CR-PLA) with intelligent reflecting surfaces (IRSs). Drawing from prior work, we establish bounds on performance metrics, such as probabilities of false alarm and missed detection, using Kullback-Leibler (KL) divergence. Leveraging prior results in [1], we extend the analysis to the CR-PLA scenario with IRS. We derive the optimal attack strategy to minimize the divergence between authentic and forged signals when the attacker has either partial or no knowledge of the legitimate cascade channel. We evaluate the attack performance under different conditions, by varying the correlation between Eve's observation and the legitimate cascade channel, or the SNR at the legitimate receiver.