On the performance of interference alignment under weak interference conditions

The notion of interference alignment (IA) has recently paved the way for exploiting interference so as to increase sum-rate capacity of wireless interference-limited systems. However, the vast majority of IA analysis of the K-user interference channel considers rich multipath conditions of equal power for all wireless links, which is not always a valid assumption for emerging wireless cellular networks. In this paper, we investigate the performance of the interference channel over weak-to-moderate interference multiple-input multiple-output (MIMO) channels. Based on the intuition that, in perfect line-of-sight (LOS) channels, with no interference between them, the sum rate scales linearly with the number of communication pairs (K), e.g. utilizing per-link MIMO water-filling technique, we explore how the sum rate scales in the case of weak-to-moderate interference across the channel pairs. Using maximum-SINR-based IA in the spatial domain, our performance evaluation results show that, indeed, in certain cases of low cross-channel interference, the IA solution corresponds to a sum rate scaling that ranges between K/2 and K, thus bridging the gap between the well-known equal-power Rayleigh case, i.e. strong interference, and the ideal scenario of non-interfered parallel MIMO channels.