Joint Design of Fronthaul and Access Links in Massive MIMO Based CRANs

ABSTRACT In this paper, we address the challenge of limited channel capacity of wireless fronthaul links in cloud radio access networks (CRANs). A novel architecture is proposed in which the base band unit (BBU) is equipped with a massive multiple‐input multiple‐output (MIMO) antenna array to communicate with a set of remote radio heads (RRHs) that serve multiple user equipments (UEs). Moreover, an optimal signal quantization strategy is also designed at the BBU to accommodate for the limited fronthaul capacity. To this end, a joint optimization of power allocation at the BBU, quantization noise covariance, and RRH beamforming vectors is formulated via maximizing the achievable sum‐rate subject to power constraints at the BBU and RRHs. The resulting highly nonconvex problem is reformulated as a semidefinite relaxation (SDR) problem by using the difference of convex (DC) programming approach. It is also deduced analytically that the solution yielded by this SDR problem is always of rank one. Numerical results confirm the potential of the proposed system in improving the capacity of the CRAN systems with wireless fronthauling. In particular, the proposed design outperforms two proposed benchmark schemes.