Knowledge-Driven Deep Learning Based Channel Estimation for Terahertz Extra-Large MIMO Systems

Kuiyu Wang*, Zhen Gao, Yifei Zhang, Tong Qin, Rui Na, Minghui Wu, Zhongxiang Li

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Citation (Scopus)

Abstract

Accurate channel state information (CSI) is critical in wideband terahertz (THz) extra-large multiple-input multiple-output (XL-MIMO) systems. However, the existing channel estimation methods suffer from the beam squint effect. In this paper, we propose a knowledge-driven channel estimation network called generalized multiple measurement vector learned approximate message passing (GMMV-LAMP) to jointly estimate the channels, whereby the approximately same physical angles shared by different subcarriers can be exploited. In particular, we propose a frequency-dependent discrete Fourier transform (DFT) wideband redundant dictionary (WRD), so that the support shift caused by the beam squint effect is eliminated. Simulation results demonstrate that the proposed approach can effectively achieve channel estimation with only several layers at the cost of limited pilots.

Original languageEnglish
Title of host publication2023 IEEE 23rd International Conference on Communication Technology
Subtitle of host publicationAdvanced Communication and Internet of Things, ICCT 2023
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages869-874
Number of pages6
ISBN (Electronic)9798350325959
DOIs
Publication statusPublished - 2023
Event23rd IEEE International Conference on Communication Technology, ICCT 2023 - Wuxi, China
Duration: 20 Oct 202322 Oct 2023

Publication series

NameInternational Conference on Communication Technology Proceedings, ICCT
ISSN (Print)2576-7844
ISSN (Electronic)2576-7828

Conference

Conference23rd IEEE International Conference on Communication Technology, ICCT 2023
Country/TerritoryChina
CityWuxi
Period20/10/2322/10/23

Keywords

  • Knowledge-driven
  • beam squint effect
  • channel estimation
  • extra-large multiple-input multiple-output (XL-MIMO)
  • orthogonal frequency division multiplexing (OFDM)

Fingerprint

Dive into the research topics of 'Knowledge-Driven Deep Learning Based Channel Estimation for Terahertz Extra-Large MIMO Systems'. Together they form a unique fingerprint.

Cite this