Acta Optica Sinica, Volume. 42, Issue 22, 2206002(2022)

Multi-Task Optical Performance Monitoring Based on Convolutional Neural Network

Jingze Ju, Qingtian Liu, Hongzhao Li, Wei Hu, Tianxiong Feng, Lin Jiang*, and Lianshan Yan
Author Affiliations
  • Center for Information Photonics and Communications, Southwest Jiaotong University, Chengdu 611756, Sichuan , China
  • show less
    Figures & Tables(9)
    Flow chart of multi-task optical performance monitoring scheme based on convolutional neural net
    Structural diagram of MTL-CNN
    Block diagram of polarization multiplexing coherent optical fiber communication system
    Intensity profile and intensity fluctuation diagrams of PDM-QPSK/-8PSK/-8QAM/-16QAM/-32QAM/-64QAM with different OSNR. (a) Intensity profile diagram; (b) intensity fluctuation diagram
    MAE of DM-QPSK/-8QAM/-16QAM/-32QAM/-64QAM under different input features. (a) PDM-QPSK; (b) PDM-8QAM; (c) PDM-16QAM; (d) PDM-32QAM; (e) PDM-64QAM
    Comparison of MAE and RMSE under different modulation formats. (a) MAE; (b) RMSE
    Comparison of RMSE and MAE under different residual chromatic dispersions. (a) RMSE; (b) MAE
    • Table 1. Identification accuracy for PDM-QPSK/-8QAM/-16QAM/-32QAM/-64QAM under multi-task optical performance monitoring scheme

      View table

      Table 1. Identification accuracy for PDM-QPSK/-8QAM/-16QAM/-32QAM/-64QAM under multi-task optical performance monitoring scheme

      Actual modulation formatIdentified modulation format

      QPSK

      (10 dB-25 dB)

      8QAM

      (12 dB-30 dB)

      16QAM

      (12 dB-30 dB)

      32QAM

      (17 dB-33 dB)

      64QAM

      (18 dB-33 dB)

      QPSK (10 dB-25 dB)1000000
      8QAM (12 dB-30 dB)0100000
      16QAM (12 dB-30 dB)0010000
      32QAM (17 dB-33 dB)0001000
      64QAM (18 dB-33 dB)0000100
    • Table 2. Identification accuracy of different modulation schemes under different residual chromatic dispersions

      View table

      Table 2. Identification accuracy of different modulation schemes under different residual chromatic dispersions

      Modulation format

      Residual chromatic dispersion /

      (psnm-1)

      0200400600
      QPSK100100100100
      8QAM100100100100
      16QAM100100100100
      32QAM100100100100
      64QAM100100100100
    Tools

    Get Citation

    Copy Citation Text

    Jingze Ju, Qingtian Liu, Hongzhao Li, Wei Hu, Tianxiong Feng, Lin Jiang, Lianshan Yan. Multi-Task Optical Performance Monitoring Based on Convolutional Neural Network[J]. Acta Optica Sinica, 2022, 42(22): 2206002

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category: Fiber Optics and Optical Communications

    Received: Apr. 22, 2022

    Accepted: May. 24, 2022

    Published Online: Nov. 7, 2022

    The Author Email: Lin Jiang (linjiang@swjtu.edu.cn)

    DOI:10.3788/AOS202242.2206002

    Topics