A convex inner approximation technique for rank-two beamforming in multicasting relay networks

In this paper, we propose a novel scheme for single-group multicasting using a relay network. We assume a source that transmits messages via an amplify-and-forward relay network to multiple destinations. The goal is to minimize the maximum transmitted power of the relays under constraints on the signal-to-noise ratios at the destinations. To increase the degrees of freedoms in the system, the relays process two source signals jointly, using two different relay beamforming weight vectors. The Alamouti space-time block code is transmitted over two beams. Simulation results demonstrate the performance of the proposed scheme combined with a proposed sequential convex programming algorithm compared to methods of the literature and to the theoretical lower bound.