A Generalized Framework on Beamformer Design and CSI Acquisition for Single-Carrier Massive MIMO Systems in Millimeter Wave Channels

摘要

In this paper, we establish a general framework on the reduced dimensionalchannel state information (CSI) estimation and pre-beamformer design forfrequency-selective massive multiple-input multiple-output MIMO systemsemploying single-carrier (SC) modulation in time division duplex (TDD) mode byexploiting the joint angle-delay domain channel sparsity in millimeter (mm)wave frequencies. First, based on a generic subspace projection taking thejoint angle-delay power profile and user-grouping into account, the reducedrank minimum mean square error (RR-MMSE) instantaneous CSI estimator is derivedfor spatially correlated wideband MIMO channels. Second, the statisticalpre-beamformer design is considered for frequency-selective SC massive MIMOchannels. We examine the dimension reduction problem and subspace (beamspace)construction on which the RR-MMSE estimation can be realized as accurately aspossible. Finally, a spatio-temporal domain correlator type reduced rankchannel estimator, as an approximation of the RR-MMSE estimate, is obtained bycarrying out least square (LS) estimation in a proper reduced dimensionalbeamspace. It is observed that the proposed techniques show remarkablerobustness to the pilot interference (or contamination) with a significantreduction in pilot overhead.