Exploiting Markov Random Field Sparsity for Wideband Channel Estimation in Massive MIMO Systems

Xiaofeng Liu*, Wenjin Wang*, Xinrui Gong*, Xiao Fu*, Xiqi Gao*, Xiang Gen Xia

*Corresponding author for this work

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

Abstract

This paper studies the wideband channel estimation (CE) problem in massive multiple-input multiple-output (MIMO) systems employing orthogonal frequency division multiplexing (OFDM). We first develop an angle-delay domain channel model with the sampled steering vectors. Then we propose a Markov random field (MRF) prior to flexibly model the angle-delay domain channel sparsity. The constrained Bethe free energy minimization (CBFEM) framework is introduced to design a message passing algorithm for the CE problem. Finally, simulation results validate the superior performance of our proposed algorithm.

Original languageEnglish
Title of host publication2022 IEEE 14th International Conference on Wireless Communications and Signal Processing, WCSP 2022
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages1040-1045
Number of pages6
ISBN (Electronic)9781665450850
DOIs
Publication statusPublished - 2022
Externally publishedYes
Event14th IEEE International Conference on Wireless Communications and Signal Processing, WCSP 2022 - Virtual, Online, China
Duration: 1 Nov 20223 Nov 2022

Publication series

Name2022 IEEE 14th International Conference on Wireless Communications and Signal Processing, WCSP 2022

Conference

Conference14th IEEE International Conference on Wireless Communications and Signal Processing, WCSP 2022
Country/TerritoryChina
CityVirtual, Online
Period1/11/223/11/22

Keywords

  • Bethe free energy
  • Markov random field
  • Massive MIMO-OFDM
  • channel estimation
  • message passing

Fingerprint

Dive into the research topics of 'Exploiting Markov Random Field Sparsity for Wideband Channel Estimation in Massive MIMO Systems'. Together they form a unique fingerprint.

Cite this