Performance Analysis of Millimeter Wave Massive MIMO Systems in Centralized and Distributed Schemes

This paper considers downlink multi-user millimeter-wave massive multiple-input multiple-output (MIMO) systems in both centralized and distributed configurations, referred to as C-MIMO and D-MIMO, respectively. Assuming the fading channel is composite and comprised of both large-scale fading and small-scale fading, a hybrid precoding algorithm leveraging antenna array response vectors is applied into both the C-MIMO system with fully connected structure and the D-MIMO system with partially connected structure. First, the asymptotic spectral efficiency (SE) of an arbitrary user and the asymptotic average SE of the cell for the C-MIMO system are analyzed. Then, two radio access unit (RAU) selection algorithms are proposed for the D-MIMO system, based on minimal distance (D-based) and maximal signal-to-interference-plus-noise-ratio (SINR) (SINR-based), respectively. For the D-MIMO system with circular layout and D-based RAU selection algorithm, the upper bounds on the asymptotic SE of an arbitrary user and the asymptotic average SE of the cell are also investigated. Finally, numerical results are provided to assess the analytical results and evaluate the effects of the numbers of total transmit antennas and users on system performance. It is shown that, from the perspective of the cell, the D-MIMO system with D-based scheme outperforms the C-MIMO system and achieves almost alike performance compared with the SINR-based solution while requiring less complexity.