MIMO Capacity with Reduced RF Chains

Abstract

This paper presents a remarkable advance for the understanding of MIMO capacity limits with insufficient RF chains. The capacity is characterized by the maximum mutual information given any vector inputs subject to not only an average power constraint but also a sparsity constraint. It is proven that the Gaussian mixture input distribution is capacity-achieving in the high signal-to-noise-ratio (SNR) regime. The optimal mixture coefficients and the covariance matrices of the Gaussian mixtures are derived and also the corresponding achievable capacity. For the special case with a single RF chain, the optimal mixture coefficients are shown to be approximately proportional to channel gains. The capacity is approximately the maximum-ratio combining (MRC) of multiple channels in the high SNR regime. We investigate the superiority of capacity-achieving techniques: Non-Uniform Spatial Modulation (NUSM) and Non-Uniform Beamspace Modulation (NUBM) by comparing them with the best antenna/beamspace selection (BAS/BBS) and the uniform spatial/beamspace modulation (USM/UBM). The comparison results reveal that the information-guided NUSM/NUBM is optimal in the high SNR regime. Numerical results are presented to validate our analysis.