Analysis of analog and digital MRC in massive MU-MIMO systems over correlated channels

While digital multi-user (MU) maximal ratio combining (MRC) is well understood, relatively few analytical results exist for analog MU-MRC. For example, it has recently been shown that MU system performance is highly dependent on the correlation model used, but the scope is limited to digital processing. Thus, in this paper we compare the performance of analog and digital MRC, focusing on the effects of correlation. We begin by deriving the expected signal and interference powers, demonstrating that the signal-to-interference ratio decreases with correlation when users have the same correlation matrices, while it increases when their correlation matrices are different. These finite system results are then extended by deriving asymptotic signal-to-interference-and-noise ratio expressions for both analog and digital MRC for the benchmark scenarios of uncorrelated and perfectly correlated Rayleigh channels. Here, we once again demonstrate that the performance is critically dependent on the correlation scenario.