Deep Learning-Based Antenna Selection and CSI Extrapolation in Massive MIMO Systems

A critical bottleneck of massive multiple-input multiple-output (MIMO) system is the huge training overhead caused by downlink transmission, like channel estimation, beamforming and covariance observation. In this paper, we propose to use state information (CSI) a small number antennas extrapolate CSI other reduce overhead. Specifically, design deep neural network that call an antenna domain extrapolation (ADEN) can exploit correlation function among antennas. We then learning (DL) based selection (ASN) select limited for optimizing extrapolation, which conventionally type combinatorial optimization difficult solve. trickly designed constrained degradation algorithm generate differentiable approximation discrete vector such back-propagation be guaranteed. Numerical results show proposed ADEN outperforms traditional fully connected one, scheme learned ASN much better than trivially used uniform selection.