An Investigation on Energy and Spectral Efficient Robust Design of Fog Radio Access Network

Abstract

This paper investigates the design of a Cloud Radio Access Network (C-RAN) to optimize energy or spectral efficiency for imperfect Channel State Information (CSI). Due to the architecture of cloud radio access networks, the BaseBand Unit (BBU) pool can acquire CSI only indirectly via sensing at the remote radio heads. As only a digital (quantized) version is forwarded to the BBU, CSI distortion is inevitable. Moreover, we consider an evolution of the conventional C-RAN, i.e., a cache-enabled Fog Radio Access Network (F-RAN). It has been shown by many researchers, that F-RANs can not only decrease the overall delay, but reduce the costs of fronthauls and the BBU pool as well. We propose two algorithms that optimize energy or spectral efficiency of a robust F-RAN using only imperfect CSI. We further show that spectral and energy efficiency of a cache-enabled F-RAN exceeds that of C-RAN, in particular under channel uncertainty constraints.