Active Eavesdropping in Cell-Free mMIMO ISAC Systems

Abstract

Cell-free massive MIMO systems with integrated sensing and communications (CF-mMIMO-ISAC) are a promising technology for enhancing spectrum efficiency. However, these systems are vulnerable to pilot spoofing attacks (PSAs), which compromise communication confidentiality and can severely degrade sensing accuracy, particularly when the attacker is spatially separated from the sensing target. To counter this threat, this paper proposes a novel security scheme. Since PSAs create coherent signals from the attacker and legitimate users, we employ the parallel profiles with linear dependencies (PARALIND) model, which is specifically designed to handle such linearly dependent sources. The PARALIND model effectively decomposes the received signal matrix to accurately estimate the attacker’s direction-of-arrival (DoA). Based on this estimate, zero-forcing (ZF) beamforming is then applied to null the interference in the attacker’s direction. Simulation results validate that the proposed scheme successfully restores both the communication secrecy rate and the sensing performance, thereby securing the CF-mMIMO-ISAC system against active eavesdropping attacks.