Interference alignment in MIMO interference broadcast channels with imperfect CSI

Abstract

In this study, the authors propose four interference alignment (IA) schemes for multiple-input multiple-output (MIMO) interference broadcast channels (IBCs) under imperfect channel state information (CSI). They propose two new algorithms based on minimising the leakage of the interference signals under a generalised imperfect CSI model from two new points of view. The first algorithm is based on alternating minimisation algorithm. The second proposed algorithm is an improved version of minimum weighted leakage interference algorithm with a faster convergence rate. Inspired by the CSI uncertainty model, the authors present two robust IA algorithms based on exploiting stochastic properties of the CSI mismatch. The first one is based on joint signal and IA which reduces the leakage of the interference signals outside the interference subspace and forces the intended signal to fall into the orthogonal complement of the interference subspace. The second algorithm maximises the signal-to-interference-plus-noise ratio (SINR) in a per-user-based approach, while the existing maximum SINR (MSINR) algorithm in literature is a per-stream-based approach. The authors illustrate that the proposed MSINR algorithm requires less channel information and has less computational costs compared to the existing MSINR. The superiority of the proposed schemes over the existing algorithms are confirmed by simulation examples.