Deep Learning Based OFDM Physical-Layer Receiver for Extreme Mobility

Jaakko Pihlajasalo, Dani Korpi, Mikko Honkala, Janne M.J. Huttunen, Taneli Riihonen, Jukka Talvitie, Mikko A. Uusitalo, Mikko Valkama

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

113 Downloads (Pure)

Abstract

In this paper, we propose a machine learning (ML) aided physical layer receiver technique for demodulating OFDM signals that are subject to very high Doppler effects and the corresponding distortion in the received signal. Specifically, we develop a deep learning based convolutional neural network receiver system that absorbs proper two-dimensional received signal entities in time and frequency, while containing convolutional neural network layers to efficiently and reliably demodulate the bits - when properly trained - despite the substantial Doppler distortion. Representative set of numerical results is provided, in the context of 5G NR mobile communication network and corresponding base-station demodulation performance for uplink. The obtained results show that the proposed receiver system is able to clearly outperform classical LMMSE receivers that operate on subcarrier level and neglect the Doppler-induced intercarrier interference (ICI). Additionally, the proposed ML receiver has the advantage over ICI cancellation based receivers in terms of the reference signal overhead. This paper provides the description of the method and vast set of numerical results in 5G NR network context.

Original languageEnglish
Title of host publication55th Asilomar Conference on Signals, Systems and Computers, ACSSC 2021
EditorsMichael B. Matthews
PublisherIEEE
Pages395-399
Number of pages5
ISBN (Electronic)9781665458283
DOIs
Publication statusPublished - 2022
Publication typeA4 Article in conference proceedings
EventAsilomar Conference on Signals, Systems, and Computers - Pacific Grove, United States
Duration: 31 Oct 20213 Nov 2021

Publication series

NameConference Record - Asilomar Conference on Signals, Systems and Computers
Volume2021-October
ISSN (Print)1058-6393

Conference

ConferenceAsilomar Conference on Signals, Systems, and Computers
Country/TerritoryUnited States
CityPacific Grove
Period31/10/213/11/21

Keywords

  • 5G NR
  • deep learning
  • Doppler
  • intercarrier interference
  • machine learning
  • mobility
  • OFDM
  • reference signals

Publication forum classification

  • Publication forum level 1

ASJC Scopus subject areas

  • Signal Processing
  • Computer Networks and Communications

Fingerprint

Dive into the research topics of 'Deep Learning Based OFDM Physical-Layer Receiver for Extreme Mobility'. Together they form a unique fingerprint.

Cite this