Performance analysis of the Singular Value Decomposition with block-diagonalization precoding in Multi-User Massive MIMO systems

Next generation of 5G mobile communications heavily relies on exploiting the spatial degrees of freedom of the radio channels. For this reason, the idea of having base stations (BSs) with a large number of antennas is currently one of the main research topics in the wireless communications field. From a radio access point of view, this calls for searching effective methods of spatially multiplexing multiple mobile stations (MSs) - otherwise termed as users - using the same timefrequency resource. This technique is known as Multi-User Multiple-Input Multiple-Output (MU-MIMO). For the downlink direction, applying classical MIMO transmission methods for MU-MIMO schemes are impractical, as they force the MSs to perform a collaborative detection, which implies a considerable increase of control information traffic. This paper analyzes the performances of the Block Diagonalization (BD) MU-MIMO downlink technique that proposes a precoding method which mitigates the inter-user interference and thus the need of a collaborative detection at the MS side. Moreover, in the performed simulations, the Singular Value Decomposition (SVD) method has been applied over the Single-User MIMO (SU-MIMO) links that were virtually created by the BD precoding technique. The results have been compared with the simpler MU-MIMO method: ZF beamforming.