30 Tbit/s DP-16QAM传输系统中基于机器学习的非线性均衡方法

范润哲; 杨超; 罗鸣

doi:10.13756/j.gtxyj.2024.230043

您当前的位置：

首页 >

文章列表页 >

30 Tbit/s DP-16QAM传输系统中基于机器学习的非线性均衡方法

光通信系统与网络技术 | 更新时间：2024-08-05

- 30 Tbit/s DP-16QAM传输系统中基于机器学习的非线性均衡方法
- Nonlinear Equalization Method based on Machine Learning in 30 Tbit/s DP-16QAM Transmission System
- 光通信研究 2024年第4期页码：23004301-23004306
- 作者机构：
  
  1.武汉邮电科学研究院，武汉　430074
  2.中国信息通信科技集团有限公司　光通信技术和网络全国重点实验室，武汉　430074
- 作者简介：
  
  范润哲（1999－），男，河北邯郸人。硕士，主要研究方向为高速光通信系统。
  杨超，正高级工程师。E-mail：yangchao@wri.com.cn
- 基金信息：
  
  国家重点研发计划资助项目(2020YFB1805905);湖北省重点研发计划资助项目(2022BAA002)
- DOI：10.13756/j.gtxyj.2024.230043
  中图分类号： TN929
- 纸质出版日期：2024-08-10，
  
  收稿日期：2023-08-08，
  
  修回日期：2023-08-20，
- 稿件说明：
扫描看全文
范润哲，杨超，罗鸣. 30 Tbit/s DP-16QAM传输系统中基于机器学习的非线性均衡方法[J]. 光通信研究，2024(4)：230043.

Fan R Z, Yang C, Luo M. Nonlinear Equalization Method based on Machine Learning in 30 Tbit/s DP-16QAM Transmission System [J]. Study on Optical Communications, 2024(4)：230043.
范润哲，杨超，罗鸣. 30 Tbit/s DP-16QAM传输系统中基于机器学习的非线性均衡方法[J]. 光通信研究，2024(4)：230043. DOI： 10.13756/j.gtxyj.2024.230043.

Fan R Z, Yang C, Luo M. Nonlinear Equalization Method based on Machine Learning in 30 Tbit/s DP-16QAM Transmission System [J]. Study on Optical Communications, 2024(4)：230043. DOI： 10.13756/j.gtxyj.2024.230043.

摘要

【目的】

采用机器学习的方法对30 Tbit/s（60×500 Gbit/s）奈奎斯特双偏振—16进制正交幅度调制（DP-16QAM）在G.654E光纤中传输6 300 km传输系统的传输结果进行非线性均衡，降低传输误码率（BER）。

【方法】

借鉴卷积神经网络的“感受野”机制，设计“卷积核”大小，根据划分采样数据构造数据集。选取并调试合适的参数构造人工神经网络。采集在C波段的不同波长、不同光信噪比和不同入纤功率发送和接收一一对应的数据，参照经典全连接神经网络结构，根据数据集数据结构来构建神经网络，并对实部和虚部分别进行网络拟合，训练完成后将测试数据集送入网络，在这3个方面分别进行计算并与传统方法进行比对。

【结果】

文章采用两种神经网络拟合了C波段频率在191.562 5~195.987 5 THz 60个不同波长传输条件下的传输BER，在与最大似然序列估计（MLSE）算法的对比中，网络1和网络2 BER分别平均降低了23%和41%。在数据中选择了频率为193.812 5 THz的光进行入纤功率范围为14~19 dBm的计算，网络1和网络2 BER分别平均降低了32%和52%。在不同光信噪比下，网络1和网络2 BER分别平均降低了30%和57%。

【结论】

引入的两种神经网络在相干传输系统的非线性均衡中都有不俗的表现。同时，网络层数和节点数会共同影响拟合结果，一般说来，增加层数和节点数可以获取更好的拟合效果，但与之对应的便是模型参数的增多、占用空间的增大和训练时间的增长，所以在应用中应考虑实际情况，在拟合效果和模型属性中进行取舍。

Abstract

【Objective】

In this paper

machine learning method is applied to 30 Tbit/s (60 × 500 Gbit/s) Nyquist Dual Polarization-16 Quadrature Amplitude Modulation (DP-16QAM) system after 6 300 km transmission in G. 654E optical fiber. Nonlinear channel equalization is used to reduce the transmission Bit Error Rate (BER).

【Methods】

Referring to the "receptive field" mechanism of convolution neural network

the size of "convolution core" is designed

and the data set is constructed according to the divided sampling data. The artificial neural network is constructed by optimizing the parameters. The one-to-one data corresponding to the transmission and reception of different wavelengths

different optical signal-to-noise ratios

and different fiber input powers in the C-band are collected. Refer to the classic full-connection neural network structure

the neural network is constructed according to the data structure of the data set. The network fitting is carried out for the real part and the imaginary part respectively. After training stage

the test data is sent into the network

and the performances are compared with the traditional methods.

【Results】

Two kinds of neural networks are used to fit the transmission BER under 60 different wavelength transmission conditions of C band frequency from 191.562 5 to 195.987 5 THz. Compared with Maximum Likelihood Sequence Estimation (MLSE)

Network 1 has an average reduction of 23% in BER

and Network 2 has an average reduction of 41% in BER. A frequency of 193.812 5 THz is then selected for the calculation of the fiber input power ranging from 14 to 19 dBm. The average improvement in network 1 and network 2 are 32% and 52%

respectively. Under different optical signal-to-noise ratios

Network 1 has an average improvement of 30%

and Network 2 has an average improvement of 57%.

【Conclusion】

The two neural networks have excellent performance in nonlinear equalization of coherent transmission systems. At the same time

the number of network layers and nodes will jointly affect the fitting results. Increasing the number of layers and nodes can obtain better fitting results

but the corresponding parameters

training time and the required space will also increase. Therefore

in the application

the actual situation should be considered to choose between the fitting performances and the model attributes.

关键词

非线性均衡机器学习神经网络大容量传输人工智能高速光通信

Keywords

nonlinear equilibriummachine learningneural networkhigh capacity transmissionartificial intelligencehigh-speed optical communication

references

Richter T, Nolle M, Frey F, et al. Generation and Coherent Reception of 107-GBd Optical Nyquist BPSK, QPSK, and 16QAM [J]. IEEE Photonics Technology Letters, 2014, 26(9):877-880.

Raybon G, Randel S, Adamiecki A, et al. High Symbol Rate Transmission Systems for Data Rates Above 400 Gb/s Using ETDM Transmitters and Receivers [C]//2014 The European Conference on Optical Communication (ECOC). Cannes, France: IEEE, 2014: 6964240.

Zhang J W, Yu J J, Fang Y, et al. High Speed All Optical Nyquist Signal Generation and Full-band Coherent Detection[J]. Scientific Reports, 2014, 4(1):6156.

Rios-Müller R, Renaudier J, Brindel P, et al. 1-Terabit/s Net Data-rate Transceiver based on Single-carrier Nyquist-shaped 124 GBaud PDM-32QAM [C]//Optical Fiber Communications Conference and Exhibition (OFC). Los Angeles, CA, USA: IEEE, 2015: Th5B.1.

Laperle C, O’Sullivan M. Advances in High-speed DACs, ADCs, and DSP for Optical Coherent Transceivers [J]. Journal of Lightwave Technology, 2014, 32(4):629-643.

Zibar D, Carvalho L H H, Piels M, et al. Application of Machine Learning Techniques for Amplitude and Phase Noise Characterization [J]. Journal of Lightwave Technology, 2015, 33(7):1333-1343.

Zibar D, Piels M, Jones R, et al. Machine Learning Techniques in Optical Communication [J]. Journal of Lightwave Technology, 2016, 34(6):1442-1452.

Agrell E, Alvarado A, Durisi G, et al. Capacity of a Nonlinear Optical Channel with Finite Memory [J]. Journal of Lightwave Technology, 2014, 32(16):2862-2876.

Irukulapati N V, Wymeersch H, Johannisson P, et al. Stochastic Digital Backpropagation [J]. IEEE Transactions on Communications, 2014, 62(11):3956-3968.

Piels M, Olmedo M I, Xue W Q, et al. Laser Rate Equation-based Filtering for Carrier Recovery in Characterization and Communication [J]. Journal of Lightwave Technology, 2015, 33(15):3271-3279.

Yousefi M I, Kschischang F R. Information Transmission Using the Nonlinear Fourier Transform, Part III: Spectrum Modulation [J]. IEEE Transactions on Information Theory, 2014, 60(7):4346-4369.

Bülow H. Experimental Demonstration of Optical Signal Detection Using Nonlinear Fourier Transform [J]. Journal of Lightwave Technology, 2015, 33(7):1433-1439.

Prilepsky J E, Derevyanko S A, Blow K J, et al. Nonlinear Inverse Synthesis and Eigenvalue Division Multiplexing in Optical Fiber Channels [J]. Physical Review Letters, 2014, 113:013901.

Bishop C M, Nasrabadi N M. Pattern Recognition and Machine Learning [M]. New York, USA: Springer, 2006.

Yolcu U, Egrioglu E, Aladag C H. A New Linear & Nonlinear Artificial Neural Network Model for Time Series Forecasting [J]. Decision Support Systems, 2013, 54(3):1340-1347.

张旭, 罗鸣, 杨超, 等. 多波段大容量光传输系统研究[J]. 光通信研究, 2022(5):1-6.

Zhang X, Luo M, Yang C, et al. Research of Large Capacity Optical Transmission System based on S/C/L Band [J]. Study on Optical Communications, 2022(5):1-6.

贺志学, 罗鸣, 李超, 等. 30.7 Tbit/s相干光PDM 16QAM OFDM 80 km SSMF传输[J]. 光通信研究, 2012(6):1-3.

He Z X, Luo M, Li C, et al. 30.7 Tbit/s Signal Transmission over 80 km SSMF Using Coherent PDM 16QAM OFDM Modulation[J]. Study on Optical Communications, 2012(6):1-3.

浏览量

下载量

CSCD

CNKI被引量

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于机器学习的光网络监测与优化方法

基于GCNet的毫米波波束选择算法

深度学习在光纤通信系统损伤补偿中的应用

基于神经网络的编码压扩联合峰均比降低技术

光纤分布式传感系统在变电站维护中的应用