Flexible Connectivity Level Control via Rate Splitting in Fog RAN Downlink With Finite-Capacity Backhaul

In the downlink of a fog radio access network (F-RAN) system, a cloud processor (CP) sends independent messages to user equipments (UEs) through a set of edge nodes (ENs). The ENs are connected to the CP via backhaul links which are modeled as orthogonal digital links. Since the ENs in F-RAN have signal processing functionalities, it is assumed that the CP transmits UEs' messages to the ENs on the backhaul links, and the ENs perform channel encoding and linear precoding based on the received messages. Due to the finite capacity of backhaul links, it is not trivial to find the optimal connectivity level which defines the number of ENs that serve each UE. This work proposes a rate-splitting approach that resolve this problem by splitting the message for each UE into multiple submessages that are simultaneously communicated with different connectivity levels. The rate and linear precoding variables for the submessages are jointly optimized so that the proposed rate-splitting scheme always outperforms baseline schemes with fixed connectivity levels. Numerical results show the effectiveness of the proposed rate-splitting scheme.