Adaptive Compression of Massive MIMO Channel State Information with Deep Learning

TL;DR

This paper explores using deep autoencoders for lossy compression of massive MIMO system channel state information, demonstrating adaptive compression with minimal runtime complexity across different channel models.

Contribution

It introduces a deep autoencoder-based method for adaptive CSI compression in massive MIMO systems, highlighting its efficiency and practical considerations.

Findings

Autoencoders provide effective lossy compression for massive MIMO CSI.

Runtime complexity remains low regardless of compression ratio.

Adaptive compression rates are feasible based on channel conditions.

Abstract

This paper proposes the use of deep autoencoders to compress the channel information in a \review{massive} multiple input and multiple output (MIMO) system. Although autoencoders perform lossy compression, they still have adequate usefulness when applied to massive MIMO system channel state information (CSI) compression. To demonstrate their impact on the CSI, we measure the performance of the system under two different channel models for different compression ratios. We disclose a few practical considerations in using autoencoders for this propose. We show through simulation that the run-time complexity of this deep autoencoder is irrelative to the compression ratio and thus an adaptive compression rate is feasible with an optimal compression ratio depending on the channel model and the signal to noise ratio.