Low-Complexity Detection For Massive MIMO Systems Over Correlated Rician Fading

Many important research works focused in recent years on the analysis of massive multiple-input multiple-output (MIMO) systems as one of the key candidate technologies for future 5G networks. In uplink massive MIMO, the use of simple linear detection techniques at the receiver side is critical from the implementation complexity point-of-view, and contributes highly to the improvement of the transmission reliability. In this paper, the performance of linear detection techniques is investigated for an uplink scenario under a realistic correlated Rice fading channel model. We derive the achievable sum-rates for the most popular linear detectors, namely, maximum ratio combining (MRC), zero-forcing (ZF), and minimum mean square error (MMSE). The results confirm that ZF and MMSE outperform MRC. However, the two receive techniques involve matrix inversion computations, and the complexity increases with the number of antennas. To overcome this problem, we propose a low-complexity method based on the well-known Neumann series (NS), and we show quantitatively that the proposed approach yields an interesting trade-off between computational complexity and achievable performance.