.png){kind=icon}
.png){kind=icon}
Return
Joint Sparse Autoencoder Based Massive MIMO CSI Feedback
.png){kind=icon}
PRE
AI
DOI：10.1109/LCOMM.2023.3250716.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
In frequency division duplex (FDD) based massive multiple-input multiple-output (MIMO) systems, the channel state information (CSI) feedback overhead could degrade spectrum and energy efficiency. Many works have made great progress in efficient feedback. However, previous deep learning (DL)-based CSI feedback schemes considered downlink CSI only, or exploited uplink CSI in a simple way. In this letter, we propose a neural network to compress and accurately recover downlink CSI, based on FDD angle and delay reciprocity between bi-directional channels. We design a joint sparse autoencoder to learn sparse transform for compression, and introduce auxiliary uplink CSI in an explainable approach. Numerical results demonstrate that our structure can improve the reconstruction quality compared with downlink-only structure.
Keywords:
Uplink
Downlink
Transforms
Minimization
Sparse matrices
Neural networks
Massive MIMO
CSI feedback
deep learning
massive MIMO
partial reciprocity
Journal
IF:
4.4
Papers:
1.3W
Citations:
2.2W