Page 35 - 网络电信2023年11月刊
P. 35

光    通    信

                四、结语                                              [11] NAKAMURA M. Optical nonlinearity compensation based
                本文将CNN-Bi  LSTM和注意力机制相结合,用以补偿相干                       on machine learning technology[C]//Proceedings of SPIE
            光传输系统中的非线性损伤。通过注意力机制研究输入序列中                               11703,AI and Optical Data Sciences II.SPIE,2021:1170319.
            每个输入符号对于接收符号的权重分布,提出了一种基于CNN-                         [12] FERNÁNDEZ E A,SOTO A M C,GONZALEZ N G,et al.Machine
            Bi  LSTM-Attention模型的光纤非线性补偿算法,并在DP-16QAM                 learning techniques to mitigate nonlinear phase noise in
            30Gbaud相干光传输系统中进行了实验验证。实验结果表明,与                           moderate baud rate optical communication systems[M]//YI
            CNN-Bi  LSTM模型相比,在1200  km的传输距离下,该算法以降                    Yang.Intelligent System and Computing.IntechOpen,2019.
            低0.03~0.23dB的Q因子为代价,使复杂度降低了约31.6%;当                   [13] DAI Xiaoxiao,LI Xiang,LUO Ming,et al.LSTM
                                                                      networks enabled nonlinear equalization in 50-
            在相似的复杂度下进行比较时,该算法在最佳发射功率下Q因子
                                                                      Gb/s PAM-4 transmission links[J].Applied
            提高了0.43dB。因此该算法可以用于长距离相干光系统的光纤
                                                                      Optics,2019,58(22):6079-6084.
            非线性损伤补偿。
                                                                  [14] DELIGIANNIDIS S,BOGRIS A,MESARITAKIS C,et al.Compensation
                                                                      of fiber nonlinearities in digital coherent systems
                                                                      leveraging long short-term memory neural networks[J].
            参考文献                                                      Journal of Lightwave Technology,2020,38(21):5991-5999.
             [1] AGRAWAL G P.Fiber-optic communication systems[M].John
                                                                  [15] FREIRE P J,OSADCHUK Y,SPINNLER B,et al.Performance
                 Wiley&Sons,2012.
                                                                      versus complexity study of neural network equalizers in
             [2] SHARMA P,PARDESHI S,ARORA R K,et al.A review of the
                                                                      coherent optical systems[J].Journal of Lightwave Technol
                 development in the field of fiber optic communication
                                                                      ogy,2021,39(19):6085-6096.
                 systems[J].International Journal of Emerging Technology
                                                                  [16] MING Hao,CHEN Xinyu,FANG Xiansong,et al.Ultralow
                 and Advanced Engineering,2013,3(5):113-119.
                                                                      complexity long short-term memory network for
             [3] HUI Rongqing.Introduction to fiber-optic
                                                                      fiber nonlinearity mitigation in coherent optical
                 communications[M].Amsterdam:Elsevier,2019.
                                                                      communication systems[J].Journal of Lightwave Technolo
             [4] ZHANG Fan,FANG Xiansong,CHEN Xinyu.Neural network-
                                                                      gy,2022,40(8):2427-2434.
                 based fiber nonlinearity mitigation in high-speed
                                                                  [17] 刘勇江,史健芳,袁晓辉.基于注意力机制与邻域几何特征
                 coherent optical transmission systems[C]//2022 Optical
                                                                      的点云语义分割[J].电子设计工程,2023,31(5):23-27,32.
                 Fiber Communications Conference and Exhibition.San
                                                                      LIU Yongjiang,WU Jianfang,YUAN Xiaohui.Point cloud
                 Diego:IEEE,2022:1-3.
                                                                      semantic segmentation based on attention mechanism and
             [5] ZHAO Jian,LIU Yaping,XU Tianhua.Advanced DSP for
                                                                      neighborhood geometric features[J].Electronic Design
                 coherent optical fiber communication[J].Applied
                                                                      Engineering, 2023,31(5):23-27,32(in Chinese).
                 Sciences,2019,9(19):4192.
                                                                  [18] 盛煜堃,彭艳兵.基于注意力机制Bi LSTM的事件抽取方法
             [6] 蒙建宇,张洪波,张敏,等.基于IPCA-DNN算法的光纤非线性损伤
                                                                      [J].电子设计工程,2020,28(8):170-173.SHENG Yukun,PENG
                 补偿[J].光学学报,2021,41(24):2406002.MENG Jianyu,ZHANG
                                                                      Yanbing.Event extraction method of based on attention-
                 Hongbo,ZHANG Min,et al.Fiber nonlinear impairments
                                                                      based bidirectional LSTM[J]. Electronic Design
                 compensation based on IPCA-DNN algorithm[J].Acta Optica
                                                                      Engineering,2020,28(8):170-173(in Chinese).
                 Sinica,2021,41(24):2406002(in Chinese).
                                                                  [19] 江叶峰,孙少华,仇晨光,等.电网故障处置预案文本中的命名
             [7] DIEN N V,SON L T,TUAN N V,et al.Digital back-propagation
                                                                      实体识别研究[J].电力工程技术,2021,40(5):177-183.JIANG
                 optimization for high-baudrate single-channel
                                                                      Yefeng,SUN Shaohua,QIU Chenguang,et al.Named entity
                 optical fiber transmissions[J].Optics Communicatio
                                                                      recognition in power fault disposal preplan text[J].
                 ns,2021,491:126913.
                                                                      Electric Power Engineering Technology,2021,40(5):177-
             [8] WANG Chuxuan,TIAN Feng,ZHANG Qi,et al.Experimental
                                                                      183(in Chinese).
                 demonstration of low complexity nonlinear compensation
                                                                  [20] LIU Yifan,SANCHEZ V,FREIRE P J,et al.Attention-
                 algorithm based on logistic regression[J].Microwave and
                                                                      aided partial bidirectional RNN-based nonlinear
                 Optical Technology Letters,2022,64(7):1154-1161.
                                                                      equalizer in coherent optical systems[J].Optics Expre
             [9] SAAVEDRA G,LIGA G,BAYVEL P.Volterra-assisted optical
                                                                      ss,2022,30(18):32908-32923.
                 phase conjugation:a hybrid optical-digital scheme for
                                                                  [21] LI Lisha,JAMIESON K,DESALVO G,et al.Hyperband:a
                 fiber nonlinearity compensation[J].Journal of Lightwave
                                                                      novel bandit-based approach to hyperparameter
                 Technology,2019,37(10):2467-2479.
                                                                      optimization[J].The Journal of Machine Learning
             [10] SHULKIND G, NAZARATHY M. Nonlinear digital back
                                                                      Research,2017,18(1):6765-6816.
                 propagation compensator for coherent optical OFDM based
                 on factorizing the Volterra series transfer function[J].
                 Optics Express,2013,21(11):13145-13161.
            56                                        网络电信 二零二三年十二月
   30   31   32   33   34   35   36   37   38   39   40