Energy-Efficient Group-Sparse Transceiver Design for Multiuser MIMO Relaying in C-RAN

This paper addresses the problem of centralized transceiver design for multiuser MIMO amplify-and-forward (AF) relaying within a cloud radio access network (C-RAN). The aim is to optimize AF matrices of remote radio heads (RRHs) acting as relays, in order to improve the reception quality at the destinations while reducing network power consumption and feedback overhead on the fronthaul links. A two-stage method is proposed to solve this problem efficiently. The first stage relies on interference leakage minimization subject to per-relay transmit power constraints along with signal preserving constraints. To reduce the total network power, RRH selection is achieved by incorporating in the objective function a regularization term that promotes group-sparsity among the RRHs. In the second stage, to reduce feedback overhead, a different penalty term is added that induces weight-level sparsity in the AF matrix of each active RRH. For both stages, low-complexity iterative algorithms based on the alternating direction method of multipliers (ADMM) are developed to solve the corresponding regularized problems with low complexity. Extensive simulations are performed to demonstrate the explicit benefits of the proposed design method, which results in notably lower power consumption, computational complexity and weight feedback overhead than conventional approaches.