Channel Correlation Modeling and its Application to Massive MIMO Channel Feedback Reduction

Abstract

In this paper, we propose a feedback information reduction technique for massive multiple-input multiple-output (MIMO) systems. To this end, we analytically derive a covariance matrix of spatially correlated Rayleigh fading channels in closed form. The covariance matrix is expressed based on its statistics, including transmit and receive antennas' correlation factors, channel variance, and channel delay profile. The closed-form expression enables a principal component analysis (PCA)-based compression of channel state information (CSI), which allows the feedback overhead to be efficiently reduced. We also analyze the compression feedback error, bit-error-rate (BER) performance, and the spectral efficiency (SE) of the system using the PCA-based compression. Under our proposed model, numerical results verify that the PCA-based compression method significantly reduces the feedback overhead of the massive MIMO systems with marginal performance degradation from full-CSI feedback. Furthermore, we propose a new design framework by numerically showing that there exists the optimal number of transmit antennas in terms of SE for a given limited feedback amount.