DNN-Based Physical-Layer Network Coding with Low-Complexity Codec

Physical-layer network coding (PNC) is considered as a good option given its ability in doubling the throughput of a two-way relay network (TWRN). In uplink of PNC, two end nodes' packets are transmitted simultaneously, but the two channel gains are usually irrelevant, hence it is sub-optimal t...

Full description

Saved in:
Bibliographic Details
Published in:Proceedings of ... IEEE International Conference on Computer and Communications (Online) pp. 713 - 718
Main Authors: Wang, Xuesong, Xie, Xinyan, Wang, Wenhao, Qi, Wenchao, He, Yuna, Lu, Lu
Format: Conference Proceeding
Language:English
Published: IEEE 08.12.2023
Subjects:
Codecs
deep neural network
Modulation
Network coding
physical-layer network coding
Relay networks
sparse training
Throughput
Training
two-way relay network
Wireless networks
ISSN:2837-7109
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Abstract Physical-layer network coding (PNC) is considered as a good option given its ability in doubling the throughput of a two-way relay network (TWRN). In uplink of PNC, two end nodes' packets are transmitted simultaneously, but the two channel gains are usually irrelevant, hence it is sub-optimal to adopt fixed-size regular QAM-type constellations for PNC system as in conventional designs. Recently, machine learning based wireless communication system becomes an important feature for 6G wireless network. Deep neural networks (DNNs) can generate irregular constellation points accordingly and automatically to deal with the performance damage caused by the different phase shifts in both uplinks. However, DNN-based PNC systems in recent literatures have high complexity to implement, and reducing the computational complexity of DNN-based PNC architecture is of great importance for the system to be practical. In this paper, we proposed a DNN-based low-complexity codec PNC system, named DLC-PNC. Compared with the state-of-the-art DNN-based PNC schemes, DLC-PNC has a much simpler structure and requires less training parameters while having the same bit-error-rate (BER) when using 4-QAM-like modulation and having lower BER rate when using 16-QAM-like modulation when signal-to-noise ratio (SNR) is larger than 15 dB. Next, we designed a sparse training algorithm to prove that the structure of DLC-PNC can be further pruned to decrease the training parameters. Simulation results show that around 50% parameters can be pruned while maintaining the similar system performance, which further reduces the time and space complexity.
AbstractList Physical-layer network coding (PNC) is considered as a good option given its ability in doubling the throughput of a two-way relay network (TWRN). In uplink of PNC, two end nodes' packets are transmitted simultaneously, but the two channel gains are usually irrelevant, hence it is sub-optimal to adopt fixed-size regular QAM-type constellations for PNC system as in conventional designs. Recently, machine learning based wireless communication system becomes an important feature for 6G wireless network. Deep neural networks (DNNs) can generate irregular constellation points accordingly and automatically to deal with the performance damage caused by the different phase shifts in both uplinks. However, DNN-based PNC systems in recent literatures have high complexity to implement, and reducing the computational complexity of DNN-based PNC architecture is of great importance for the system to be practical. In this paper, we proposed a DNN-based low-complexity codec PNC system, named DLC-PNC. Compared with the state-of-the-art DNN-based PNC schemes, DLC-PNC has a much simpler structure and requires less training parameters while having the same bit-error-rate (BER) when using 4-QAM-like modulation and having lower BER rate when using 16-QAM-like modulation when signal-to-noise ratio (SNR) is larger than 15 dB. Next, we designed a sparse training algorithm to prove that the structure of DLC-PNC can be further pruned to decrease the training parameters. Simulation results show that around 50% parameters can be pruned while maintaining the similar system performance, which further reduces the time and space complexity.
Author Lu, Lu
Xie, Xinyan
Wang, Wenhao
He, Yuna
Wang, Xuesong
Qi, Wenchao
Author_xml – sequence: 1
  givenname: Xuesong
  surname: Wang
  fullname: Wang, Xuesong
  email: wangxuesong17@mails.ucas.ac.cn
  organization: Shanghai Electro-Mechanical Engineering Institute,Shanghai,China,201109
– sequence: 2
  givenname: Xinyan
  surname: Xie
  fullname: Xie, Xinyan
  organization: University of Chinese Academy of Sciences,Beijing,China,101408
– sequence: 3
  givenname: Wenhao
  surname: Wang
  fullname: Wang, Wenhao
  organization: East China Normal University,School of Communication and Electronic Engineering,Shanghai,China,200062
– sequence: 4
  givenname: Wenchao
  surname: Qi
  fullname: Qi, Wenchao
  organization: University of Chinese Academy of Sciences,Beijing,China,101408
– sequence: 5
  givenname: Yuna
  surname: He
  fullname: He, Yuna
  organization: University of Chinese Academy of Sciences,Beijing,China,101408
– sequence: 6
  givenname: Lu
  surname: Lu
  fullname: Lu, Lu
  email: lulu@csu.ac.cn
  organization: Chinese Academy of Sciences,Key Laboratory of Space Utilization,Beijing,China,100094
BookMark eNo1j8tOwzAUBQ0CiVLyB0jkBxx8bRzby2JelaLCAtaVY19TQ9pUSaSQv4cKWM1iNEc65-Rk1-6QkCtgBQAz10trrTTSsIIzLgpgkikp1BHJjDJaSCZAcQnHZMa1UFT9NGck6_sPxg5KaQUzsrhbreit6zHkL5upT941tHITdvkKh7HtPnPbhrR7z8c0bPKqHaltt_sGv9IwHRT6C3IaXdNj9sc5eXu4f7VPtHp-XNpFRROAGWhEf6NcVB7Ksq6jltzHEFA5r31dcoMqgi5NDKx23IEzQRoNQmBgpnQYxJxc_u4mRFzvu7R13bT-fy2-AR5mTjI
ContentType Conference Proceeding
DBID 6IE
6IL
CBEJK
RIE
RIL
DOI 10.1109/ICCC59590.2023.10507537
DatabaseName IEEE Electronic Library (IEL) Conference Proceedings
IEEE Xplore POP ALL
IEEE Xplore All Conference Proceedings
IEEE Electronic Library (IEL)
IEEE Proceedings Order Plans (POP All) 1998-Present
DatabaseTitleList
Database_xml – sequence: 1
  dbid: RIE
  name: IEEE Electronic Library (IEL)
  url: https://ieeexplore.ieee.org/
  sourceTypes: Publisher
DeliveryMethod fulltext_linktorsrc
EISBN 9798350317251
EISSN 2837-7109
EndPage 718
ExternalDocumentID 10507537
Genre orig-research
GroupedDBID 6IE
6IF
6IK
6IL
6IN
AAJGR
ABLEC
ADZIZ
ALMA_UNASSIGNED_HOLDINGS
BEFXN
BFFAM
BGNUA
BKEBE
BPEOZ
CBEJK
CHZPO
IEGSK
IPLJI
M43
OCL
RIE
RIL
ID FETCH-LOGICAL-i119t-fec47af7c166bbf852cfdde7ac8cb629e7f1869fd0ba2a1a9d598133ed096aed3
IEDL.DBID RIE
IngestDate Tue May 06 03:32:21 EDT 2025
IsPeerReviewed false
IsScholarly false
Language English
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-i119t-fec47af7c166bbf852cfdde7ac8cb629e7f1869fd0ba2a1a9d598133ed096aed3
PageCount 6
ParticipantIDs ieee_primary_10507537
PublicationCentury 2000
PublicationDate 2023-Dec.-8
PublicationDateYYYYMMDD 2023-12-08
PublicationDate_xml – month: 12
  year: 2023
  text: 2023-Dec.-8
  day: 08
PublicationDecade 2020
PublicationTitle Proceedings of ... IEEE International Conference on Computer and Communications (Online)
PublicationTitleAbbrev ICCC
PublicationYear 2023
Publisher IEEE
Publisher_xml – name: IEEE
SSID ssj0003177871
Score 1.8608873
Snippet Physical-layer network coding (PNC) is considered as a good option given its ability in doubling the throughput of a two-way relay network (TWRN). In uplink of...
SourceID ieee
SourceType Publisher
StartPage 713
SubjectTerms Codecs
deep neural network
Modulation
Network coding
physical-layer network coding
Relay networks
sparse training
Throughput
Training
two-way relay network
Wireless networks
Title DNN-Based Physical-Layer Network Coding with Low-Complexity Codec
URI https://ieeexplore.ieee.org/document/10507537
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1LSwMxEA62ePCkYsU3OXhN3Wx28zjqalEoSw8KvZU8JtBLV-rW4r83SbcVDx68hTwgmWQyj8zkQ-iWZ8YIz0vCOAApqKFEFqACu2uXO-YiJmMCmxB1LadTNemS1VMuDACk4DMYxmJ6y3eNXUVXWeDwoL2UTPRQTwi-SdbaOVSCIAyHj3YxXDRTdy9VVZWqVNkwYoQPt6N_4agkMTI6_OcEjtDgJyEPT3ai5hjtweIE3T_WNXkIUsjhSUdtMtZBhcb1JrYbV03sjqOvFY-bNYnMHz_AbL9iE9gBehs9vVbPpENEIHNKVUs82EJoLyzl3Bgvy9z6cD8JbaU1PFcgfISY8i4zOtdUK1cqGaxQcMFS0eDYKeovmgWcIcwUDaaE9Bl3rDDcKQlKKm4LCUJ4x87RIK5_9r759GK2XfrFH_WX6CBSOUV6yCvUb5cruEb79rOdfyxv0lZ9A7balEQ
linkProvider IEEE
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV07T8MwELagIMEEiCLeZGB1iZP4NUKgakWIOhSpW-XHWerSoJKC-PfYaVrEwMBm-SHZZ5_v4Tt_CN2yWGvuGMUpA8AZ0QSLDKRnd2UTm9qAydiATfCyFJOJHLXJ6k0uDAA0wWfQC8XmLd9WZhlcZZ7DvfZCU76NdmiWJfEqXWvjUvGi0B8_0kZxkVjeDfM8p5LKuBdQwnvr8b-QVBpB0j_45xQOUfcnJS8abYTNEdqC-TG6fyxL_ODlkI1GLb1xobwSHZWr6O4or0L3KHhbo6L6xIH9wxeY9VdoAtNFr_2ncT7ALSYCnhEia-zAZFw5bghjWjtBE-P8DcWVEUazRAJ3AWTK2VirRBElLZXC26Fgva2iwKYnqDOv5nCKolQSb0wIFzObZppZKUAKyUwmgHNn0zPUDeufvq2-vZiul37-R_0N2huMX4ppMSyfL9B-oHgT9yEuUadeLOEK7ZqPeva-uG627RufjZeL
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Abook&rft.genre=proceeding&rft.title=Proceedings+of+...+IEEE+International+Conference+on+Computer+and+Communications+%28Online%29&rft.atitle=DNN-Based+Physical-Layer+Network+Coding+with+Low-Complexity+Codec&rft.au=Wang%2C+Xuesong&rft.au=Xie%2C+Xinyan&rft.au=Wang%2C+Wenhao&rft.au=Qi%2C+Wenchao&rft.date=2023-12-08&rft.pub=IEEE&rft.eissn=2837-7109&rft.spage=713&rft.epage=718&rft_id=info:doi/10.1109%2FICCC59590.2023.10507537&rft.externalDocID=10507537