Feature based modulation classification for overlapped signals

Yizhou Jiang, Sai Huang, Yixin Zhang, Zhiyong Feng, Di Zhang, Celimuge Wu

Research output: Contribution to journalArticle

Abstract

This letter proposes a novel modulation classification method for overlapped sources named LRGP involving multinomial logistic regression (MLR) and multi-gene genetic programming (MGGP). MGGP based feature engineering is conducted to transform the cumulants of the received signals into highly discriminative features and a MLR based classifier is trained to identify the combination of the modulation formats of the overlapped sources instead of signal separation. Extensive simulations demonstrate that LRGP yields superior performance compared with existing methods.

Original languageEnglish
Pages (from-to)1123-1126
Number of pages4
JournalIEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences
VolumeE101A
Issue number7
DOIs
Publication statusPublished - 2018 Jul 1
Externally publishedYes

Fingerprint

Genetic programming
Logistic Regression
Genetic Programming
Logistics
Modulation
Genes
Gene
Cumulants
Classifiers
Classifier
Transform
Engineering
Demonstrate
Simulation

Keywords

  • Cumulant
  • Modulation classification
  • Multi-gene genetic programming
  • Multinomial logistic regression
  • Overlapped sources

ASJC Scopus subject areas

  • Signal Processing
  • Computer Graphics and Computer-Aided Design
  • Electrical and Electronic Engineering
  • Applied Mathematics

Cite this

Feature based modulation classification for overlapped signals. / Jiang, Yizhou; Huang, Sai; Zhang, Yixin; Feng, Zhiyong; Zhang, Di; Wu, Celimuge.

In: IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences, Vol. E101A, No. 7, 01.07.2018, p. 1123-1126.

Research output: Contribution to journalArticle

Jiang, Yizhou ; Huang, Sai ; Zhang, Yixin ; Feng, Zhiyong ; Zhang, Di ; Wu, Celimuge. / Feature based modulation classification for overlapped signals. In: IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences. 2018 ; Vol. E101A, No. 7. pp. 1123-1126.
@article{0788982e1fdf46fe804d8c320cbe243e,
title = "Feature based modulation classification for overlapped signals",
abstract = "This letter proposes a novel modulation classification method for overlapped sources named LRGP involving multinomial logistic regression (MLR) and multi-gene genetic programming (MGGP). MGGP based feature engineering is conducted to transform the cumulants of the received signals into highly discriminative features and a MLR based classifier is trained to identify the combination of the modulation formats of the overlapped sources instead of signal separation. Extensive simulations demonstrate that LRGP yields superior performance compared with existing methods.",
keywords = "Cumulant, Modulation classification, Multi-gene genetic programming, Multinomial logistic regression, Overlapped sources",
author = "Yizhou Jiang and Sai Huang and Yixin Zhang and Zhiyong Feng and Di Zhang and Celimuge Wu",
year = "2018",
month = "7",
day = "1",
doi = "10.1587/transfun.E101.A.1123",
language = "English",
volume = "E101A",
pages = "1123--1126",
journal = "IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences",
issn = "0916-8508",
publisher = "Maruzen Co., Ltd/Maruzen Kabushikikaisha",
number = "7",

}

TY - JOUR

T1 - Feature based modulation classification for overlapped signals

AU - Jiang, Yizhou

AU - Huang, Sai

AU - Zhang, Yixin

AU - Feng, Zhiyong

AU - Zhang, Di

AU - Wu, Celimuge

PY - 2018/7/1

Y1 - 2018/7/1

N2 - This letter proposes a novel modulation classification method for overlapped sources named LRGP involving multinomial logistic regression (MLR) and multi-gene genetic programming (MGGP). MGGP based feature engineering is conducted to transform the cumulants of the received signals into highly discriminative features and a MLR based classifier is trained to identify the combination of the modulation formats of the overlapped sources instead of signal separation. Extensive simulations demonstrate that LRGP yields superior performance compared with existing methods.

AB - This letter proposes a novel modulation classification method for overlapped sources named LRGP involving multinomial logistic regression (MLR) and multi-gene genetic programming (MGGP). MGGP based feature engineering is conducted to transform the cumulants of the received signals into highly discriminative features and a MLR based classifier is trained to identify the combination of the modulation formats of the overlapped sources instead of signal separation. Extensive simulations demonstrate that LRGP yields superior performance compared with existing methods.

KW - Cumulant

KW - Modulation classification

KW - Multi-gene genetic programming

KW - Multinomial logistic regression

KW - Overlapped sources

UR - http://www.scopus.com/inward/record.url?scp=85049388344&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85049388344&partnerID=8YFLogxK

U2 - 10.1587/transfun.E101.A.1123

DO - 10.1587/transfun.E101.A.1123

M3 - Article

AN - SCOPUS:85049388344

VL - E101A

SP - 1123

EP - 1126

JO - IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences

JF - IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences

SN - 0916-8508

IS - 7

ER -