부산대학교 도서관

Cell-free massive multiple-input-multiple-out (CF-mMIMO) systems provide limited coverage because of expensive wired fronthaul between access points (APs) and central processing unit (CPU). To address this challenge, we propose a novel framework where an unmanned aerial vehicle (UAV), acting as an aerial AP, works coherently with the ground APs to expand the coverage of conventional CF-mMIMO system. Due to limited capacity of the wired and wireless fronthaul, the compress-and-forward scheme is introduced. To avoid additional use of spectrum resources, the wireless fronthaul shares the total bandwidth with the radio access networks. Considering limited power supply of the UAV and for the goal of green communications, we formulate a weighted sum power minimization problem to jointly optimize downlink beamforming and fronthaul compression, as well as UAV placement. Then, we propose a two-timescale optimization framework in which the downlink beamforming and fronthaul compression are optimized in each short timescale based on instantaneous channel state information (CSI), whereas the UAV placement is optimized in each long timescale based on statistical CSI. Specifically, uplink-downlink duality and semidefinite relaxation (SDR) based alternating optimization techniques are introduced to find solutions to the short timescale issue, while an iterative algorithm based on successive convex approximation and SDR methods is proposed to find solutions to the long timescale issue. Simulation results demonstrate that the proposed two-timescale algorithm outperforms the baseline algorithms.