Inter-cluster design of precoding and fronthaul compression for cloud radio access networks

This work studies the joint design of precoding and fronthaul compression for the downlink of multi-cluster cloud radio access networks (C-RANs). In these systems, each cluster of radio units (RUs) is managed by a different control unit (CU) through fronthaul links of limited capacity. Each CU performs baseband processing, including channel coding and linear precoding, for all the connected RUs, and then compresses the precoded signals for transmission to the RUs over the corresponding fronthaul links. In a recent work, it was shown that joint, rather than conventional separate, compression of the signals of different RUs leads to significant performance gains. However, the approach therein was limited to the intra-cluster optimization of precoding and joint compression, hence leading to significant performance loss in the presence of mutually interfering clusters. This work proposes instead to perform the inter-cluster optimization of precoding and joint compression. Specifically, the problem of maximizing the weighted sum-rate maximization across multiple clusters is tackled by deriving an iterative algorithm that converges to a locally optimal point of the problem. From numerical results, it is confirmed that the proposed scheme significantly outperforms the existing approach based on intra-cluster optimization.