Relay-Aided Downlink Massive MIMO NOMA With Estimated CSI

A multiple relay-aided downlink (DL) multi-user massive multiple-input multiple-output (MIMO) non-orthogonal multiple access (NOMA) system is investigated to support simultaneous connectivity to cell-edge users. The estimated uplink channel state information (CSI) is used for precoder designs, and statistical DL CSI is adopted for successive interference cancellation (SIC) at the users. The achievable rates are derived, and thereby, the deleterious effects of CSI uncertainty, imperfect SIC, inter-relay interference and intra-cluster pilot contamination are quantified. Fundamental trade-offs between the achievable sum rate and system scalability are explored via two pilot assignment schemes. The performance of the proposed system is compared with that of the orthogonal multiple access (OMA) counterpart. We reveal that the latter performs better in low-user regime because the former's achievable rates are hindered by intra-cluster pilot contamination and imperfect SIC.However, the proposed model provides better system scalability and outperforms OMA in the large-user regime. The performance of zero-forcing (ZF)/null-space (NS) precoding is compared with maximum ratio transmission (MRT). We show that NS-based ZF precoding can boost the sum rate by mitigating inter-cluster interference at the expense of a higher computational complexity than that of MRT.