Functional Split of Zero-Forcing Based Massive MIMO for Fronthaul Load Reduction

Abstract

The evolution towards centralized radio access network (C-RAN) for 5G is driven by the need for improved network performance and reduced total-cost-of-ownership (TCO). In C-RAN, physically separated remote radio units (RRUs) and baseband units (BBUs) are connected via fronthaul (FH) links, which are capacity constrained. When base-station antennas are upgraded to support massive MIMO techniques, an intra PHY split between RRU and BBU becomes favorable to avoid the dramatic rate growth on the FH links. In this paper, we present a new uplink functional split alternative that decomposes the massive MIMO processing, on the basis of zero-forcing (ZF) equalization, into two phases: dimension-reduction phase and interference cancellation phase, which are implemented in RRU and BBU respectively. Compared with the traditional C-RAN architecture having all baseband processing in the BBU, the new scheme largely alleviates the FH load by reducing the number of FH streams to be equal to the number of MIMO user layers. Compared with an existing split that places all MIMO processing in RRU, the new scheme reduces the RRU complexity without compromising the post-processing performances.