Collusive Eavesdropping in Interference Alignment Based Wireless Networks

Interference alignment (IA) can be secure due to the fact that the received signal of a targeted user at the eavesdropper may be embedded by interference from other concurrent users. However, when some malicious users inside the network cooperate to eavesdrop one specific user, the network will not be secure any more. Thus, we focus on eavesdropping attacks, and propose a novel collusive eavesdropping scheme (CES) in a K-user IA-based network. In this scheme, one user is eavesdropped on by an eavesdropper with the aid of the other (K - 2) users. To perform passive eavesdropping without being noticed by the targeted user, the precoding and decoding matrices of the eavesdropper and its cooperators are redesigned, and some of the cooperators sacrifice their own quality of transmission to help the eavesdropper meet the feasibility condition. Therefore, the feasibility condition of CES is derived, based on which the minimal number of low-quality cooperators and the maximal number of receiving antennas at each user are obtained. The received power of eavesdropping is analyzed with different numbers of antennas at each receiver, which also affects the eavesdropping performance. Extensive simulation results are provided to show the effectiveness of CES.