2020 Volume 9 Issue 5 Pages 146-151
Multiuser multiple-input multiple-output (MU-MIMO) downlink technology over massive MIMO is a key enabler for 5G systems and beyond. However, its performance may deteriorate when users are spatially correlated. Distributed antenna deployment (DAD) and nonlinear precoding (NLP) have been studied as countermeasures for this issue. In this paper, through simulations of 28 GHz-band massive MIMO downlink transmission in an ultra-dense urban scenario, it is revealed that, rather than DAD, concentrated antenna deployment (CAD) is more robust against channel transition issues up to walking speed. It is also found that combining CAD with NLP provided the best throughput performance in the evaluated scenario.