Scalable Massive Random Access in C- RAN with Fronthaul Limitations

Abstract

Massive connectivity for massive machine-type communications (mMTC) is key enabler for many 5G and Intternet-of-Things (IoT) applications. In mMTC scenarios, a large number of low-complexity, low-rate devices transmit sporadically over shared scarce communication resources. Reliability and low-latency with short packet transmissions are major design criteria that require integration of novel network architecture solutions and multiple access methods. This study addresses the challenge of massive connectivity in the context of a Cloud Radio Access Network (C-RAN), characterized by a hierarchical structure in which part of the processing functionalities of the Radio Units (RUs) are migrated to a centralized cloud processor or Central Unit (CU). A major architectural constraint of C-RANs is imposed by capacity-limited fronthaul links connecting the RUs with the CU. In this study, the performance of a transmission scheme for massive connectivity is investigated under this architectural constraint. In particular, a non-coherent, grant-free transmission scheme is proposed where the encoding is performed based on a Gabor frame structure. The fronthaul processing is a particular instance of Detect-and-Forward (DtF), where local detection at the RUs is performed using a one-step thresholding procedure, and the local estimates are then merged at the central processor via the capacity-limited fronthaul links. Numerical results illustrate the potential of the proposed scheme to support massive, reliable random access with short messages.