Research on Blind Recognition Algorithm of Channel Coding Based on One-Dimensional Convolutional Neural Network Under the Low SNR Regime

Pan Deng*, Tianqi Zhang, Baoze Ma, Zeliang An

*Kontaktforfatter

Publikation: Bidrag til tidsskriftTidsskriftartikelForskningpeer review

Abstract

To solve the problem of blind identification of channel coding in signal interception and intelligent communication systems, a blind recognition algorithm of channel coding based on a convolutional neural network (CNN) is proposed. To deal better with the input characteristics of one-dimensional soft decision sequences under the low SNR regime, we design and construct a low-complexity one-dimensional CNN classifier. Firstly, the dataset of five types of channel codes is generated, including linear block codes, Turbo codes, convolutional codes, LDPC codes and Polar codes. Secondly, the one-dimensional CNN classifier is initialized and imported into the database. Thirdly, the training dataset.mat file generated in matlab is imported into the convolutional neural network, and the dataset of five types of channel codes is labeled. Finally, a one-dimensional convolutional neural network model is established, and the dataset in the model is compiled and trained, the training data tags in the model are predicted and the confusion matrix is saved. The simulation results show that the proposed algorithm expands the recognition range of existing types of channel codes, and for the first time completes the blind recognition of five types of channel codes. Compared with the traditional algorithm LSTM (Long Short Term Memory), the algorithm proposed in this paper has lower complexity, the recognition rate can be kept 92% when the SNR is 0 dB, and the performance is improved by 24%.

OriginalsprogEngelsk
TidsskriftNeural Processing Letters
Vol/bind55
Sider (fra-til)5653-5673
Antal sider21
ISSN1370-4621
DOI
StatusUdgivet - okt. 2023
Udgivet eksterntJa

Bibliografisk note

Publisher Copyright:
© 2023, The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature.

Fingeraftryk

Dyk ned i forskningsemnerne om 'Research on Blind Recognition Algorithm of Channel Coding Based on One-Dimensional Convolutional Neural Network Under the Low SNR Regime'. Sammen danner de et unikt fingeraftryk.

Citationsformater