MIMO Broadcast Channels with Full-duplex feedback
Master of Science
In this thesis, I utilize full-duplex radios to increase the downlink spectral efficiency of MIMO broadcast channels by proposing a sequential beamforming strategy. In current half-duplex multiuser MIMO systems, non-trivial time resources is devoted into the collection of channel state information (CSI). In sequential beamforming strategy, instead of waiting for CSI from all users being collected, base station immediately begins to serves users when only partial CSI is available. Such operation, however, brings new inter-node interference. I quantify the impact of this unique new interference and the associated tradeoff in the design of the training control channel with closed-form bounds and simulation results in both finite SNR and high SNR regime. With experimental results collected in both indoor and outdoor environment, I further demonstrate that sequential beamforming achieves significant spectral efficiency improvement for systems with both open-loop training and closed-loop training.