Transmit-receive beamforming optimization for full-duplex cloud radio access networks

We consider a cloud radio access network (CRAN) with full duplex (FD) remote radio heads (RRHs) and half duplex mobile users. Compared with half duplex RRHs, though FD-RRHs can simultaneously transmit and receive data streams, they also suffer from new interference sources such as self-interference and inter-RRH interference. With FDRRHs, the downlink mobile users (DMUs) are also interfered by signals from the uplink mobile users (UMUs). To mitigate the interference aforementioned, new beamforming designs are required for downlink transmission and uplink reception at the FD-RRHs. We propose to minimize the sum power of CRAN by optimizing the beamformers of FD-RRHs and power control of UMUs, under quality of service constraints for both DMUs and UMUs. While the considered problem is not convex due to the new interference sources, we can solve it by second-ordercone-program (SOCP) based alternating optimization (AO) with guaranteed convergence to the KKT point. Moreover, we show that there still holds an interesting uplink-downlink duality in our problem. This duality is exploited to develop another AO solver with the same performance. The duality-based AO solver has much lower complexity than the SOCP-based one, and the simulation results show that both AO solvers yields to smaller sum power compared with the half duplex CRAN.