Massive MIMO operation in partially centralized cloud radio access networks

Massive multiple-input multiple-output (MIMO) is considered as one of the key technologies in next generation cellular systems due to its higher multiplexing gain and energy efficiency. However, in a cloud radio access network (C-RAN) with massive MIMO, operation of many antennas incurs a huge amount of digital sampled data to be transported over the fronthaul link between a baseband unit (BBU) pool and each remote radio head (RRH). Accordingly, the bandwidth requirement for the fronthaul link increases in proportion to the number of antennas at the RRH, and it incurs enormous fronthaul link cost to operators. In this paper, we propose a new C-RAN architecture called massive MIMO partially-centralized C-RAN (MPC-RAN), which is designed for a more efficient use of fronthaul resources in massive MIMO operation. Differently from a conventional fully-centralized C-RAN (FC-RAN) where most L1/L2/L3 functions are performed at the BBU pool, some beamforming-related functions in the proposed MPC-RAN are performed at RRHs. We further provide an MPC-RAN operation strategy that aims to efficiently exploit the proposed C-RAN structure in massive MIMO systems. Specifically, we formulate a wireless sum-rate maximization problem for constrained fronthaul capacity and provide a heuristic algorithm that targets at obtaining optimal beamforming configuration and bandwidth allocation. The simulation results confirm that our proposed architecture requires less fronthaul link capacity compared to FC-RAN for the same transmission configuration, and our operation strategy allows for flexible utilization of capacity-limited fronthaul resources.