Experimental demonstration of low complexity nonlinear compensation algorithm based on logistic regression

Chuxuan Wang; Feng Tian; Qi Zhang; Ran Gao; Yongjun Wang; Qinghua Tian; Xishuo Wang; Xiaolong Pan; Xiangjun Xin

doi:10.1002/mop.33252

Experimental demonstration of low complexity nonlinear compensation algorithm based on logistic regression

Chuxuan Wang, Feng Tian^*, Qi Zhang, Ran Gao, Yongjun Wang, Qinghua Tian, Xishuo Wang, Xiaolong Pan, Xiangjun Xin

^*Corresponding author for this work

School of Information and Electronics

Research output: Contribution to journal › Article › peer-review

1 Citation (Scopus)

Abstract

In this paper, a novel algorithm based on logistic regression digital backpropagation (LR-DBP) is proposed to mitigate the effect of dispersion and nonlinearity in a long-distance optical transmission system. The LR-DBP can evaluate nonlinear effects by estimating the cross-entropy error and adjusting the structure of DBP. The experiment is setup to demonstrate an 80 Gbit/s 16QAM signal transmitted across 1000 km single-mode fiber. The results show that LR-DBP achieves a 1.8 dB launch power improvement over the dispersion compensation algorithm and a 0.6 dB launch power gain with DBP. The proposed algorithm could obtain with 19.2% computational complexity decrease and 19.6% central processing unit (CPU) running time decrease by calculating the average numbers of real multiplications and CPU running times.

Original language	English
Pages (from-to)	1154-1161
Number of pages	8
Journal	Microwave and Optical Technology Letters
Volume	64
Issue number	7
DOIs	https://doi.org/10.1002/mop.33252
Publication status	Published - Jul 2022

Keywords

logistic regression
machine learning
nonlinearity
optical fiber communication

Access to Document

10.1002/mop.33252

Cite this

@article{dadf3dbffccb4b4ba3a57454576c99ab,

title = "Experimental demonstration of low complexity nonlinear compensation algorithm based on logistic regression",

abstract = "In this paper, a novel algorithm based on logistic regression digital backpropagation (LR-DBP) is proposed to mitigate the effect of dispersion and nonlinearity in a long-distance optical transmission system. The LR-DBP can evaluate nonlinear effects by estimating the cross-entropy error and adjusting the structure of DBP. The experiment is setup to demonstrate an 80 Gbit/s 16QAM signal transmitted across 1000 km single-mode fiber. The results show that LR-DBP achieves a 1.8 dB launch power improvement over the dispersion compensation algorithm and a 0.6 dB launch power gain with DBP. The proposed algorithm could obtain with 19.2% computational complexity decrease and 19.6% central processing unit (CPU) running time decrease by calculating the average numbers of real multiplications and CPU running times.",

keywords = "logistic regression, machine learning, nonlinearity, optical fiber communication",

author = "Chuxuan Wang and Feng Tian and Qi Zhang and Ran Gao and Yongjun Wang and Qinghua Tian and Xishuo Wang and Xiaolong Pan and Xiangjun Xin",

note = "Publisher Copyright: {\textcopyright} 2022 Wiley Periodicals LLC.",

year = "2022",

month = jul,

doi = "10.1002/mop.33252",

language = "English",

volume = "64",

pages = "1154--1161",

journal = "Microwave and Optical Technology Letters",

issn = "0895-2477",

publisher = "John Wiley & Sons Inc.",

number = "7",

}

TY - JOUR

T1 - Experimental demonstration of low complexity nonlinear compensation algorithm based on logistic regression

AU - Wang, Chuxuan

AU - Tian, Feng

AU - Zhang, Qi

AU - Gao, Ran

AU - Wang, Yongjun

AU - Tian, Qinghua

AU - Wang, Xishuo

AU - Pan, Xiaolong

AU - Xin, Xiangjun

PY - 2022/7

Y1 - 2022/7

N2 - In this paper, a novel algorithm based on logistic regression digital backpropagation (LR-DBP) is proposed to mitigate the effect of dispersion and nonlinearity in a long-distance optical transmission system. The LR-DBP can evaluate nonlinear effects by estimating the cross-entropy error and adjusting the structure of DBP. The experiment is setup to demonstrate an 80 Gbit/s 16QAM signal transmitted across 1000 km single-mode fiber. The results show that LR-DBP achieves a 1.8 dB launch power improvement over the dispersion compensation algorithm and a 0.6 dB launch power gain with DBP. The proposed algorithm could obtain with 19.2% computational complexity decrease and 19.6% central processing unit (CPU) running time decrease by calculating the average numbers of real multiplications and CPU running times.

AB - In this paper, a novel algorithm based on logistic regression digital backpropagation (LR-DBP) is proposed to mitigate the effect of dispersion and nonlinearity in a long-distance optical transmission system. The LR-DBP can evaluate nonlinear effects by estimating the cross-entropy error and adjusting the structure of DBP. The experiment is setup to demonstrate an 80 Gbit/s 16QAM signal transmitted across 1000 km single-mode fiber. The results show that LR-DBP achieves a 1.8 dB launch power improvement over the dispersion compensation algorithm and a 0.6 dB launch power gain with DBP. The proposed algorithm could obtain with 19.2% computational complexity decrease and 19.6% central processing unit (CPU) running time decrease by calculating the average numbers of real multiplications and CPU running times.

KW - logistic regression

KW - machine learning

KW - nonlinearity

KW - optical fiber communication

UR - http://www.scopus.com/inward/record.url?scp=85127410750&partnerID=8YFLogxK

U2 - 10.1002/mop.33252

DO - 10.1002/mop.33252

M3 - Article

AN - SCOPUS:85127410750

SN - 0895-2477

VL - 64

SP - 1154

EP - 1161

JO - Microwave and Optical Technology Letters

JF - Microwave and Optical Technology Letters

IS - 7

ER -

Experimental demonstration of low complexity nonlinear compensation algorithm based on logistic regression

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this