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

作　　者：范润哲 杨超 罗鸣 FAN Runzhe;YANG Chao;LUO Ming(Wuhan Research Institute of Posts and Telecommunications,Wuhan 430074,China;National Key Laboratory of Optical Communication Technologies and Networks,China Information Communication Technologies Group Corporation,Wuhan 430074,China)

机构地区：[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

出　　处：《光通信研究》2024年第4期7-12,共6页Study on Optical Communications

基　　金：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)。

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6300 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191.5625~195.9875 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193.8125 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。【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 6300 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 191.5625 to 195.9875 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 193.8125 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 th

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 

分 类 号：TN929[电子电信—通信与信息系统]