Acta Photonica Sinica, Volume. 53, Issue 7, 0706001(2024)

Deep Learning Aided Signal Detection Algorithm Experimental Research for Underwater Optical Communication

Pengfei YE1...3, Peng ZHANG2,*, Hao YU2, Shuang HE2, Dongsheng TIAN2, Yuanxin WANG2,4 and Shoufeng TONG2 |Show fewer author(s)
Author Affiliations
  • 1School of Electronic Information Engineering, Changchun University of Science and Technology, Changchun 130012, China
  • 2School of Electro-optical Engineering, Changchun University of Science and Technology, Changchun 130012, China
  • 3Zhongshan Research Institute, Changchun University of Science and Technology, Zhongshan 528400,China
  • 4College of Communication Engineering, Jilin University, Changchun 130012, China
  • show less
    Figures & Tables(12)
    Neural network model
    Experimental plan for indoor underwater optical communication
    Indoor UWOC experimental platform
    Neural network structure
    Waveform test
    PDF of water tank channel
    Methods for obtaining datasets and training models
    Simulation of BER in three different channels
    BER in flow channel
    BER for turbid flow channel
    BER at different rates
    • Table 1. Neural network training parameters

      View table
      View in Article

      Table 1. Neural network training parameters

      ParameterNumber
      Epochs200
      Mini-batchs64
      Dropout ratio0.1
      Ratio of validation data0.1
      OptimizerAdam
      Initial learning rate0.001
    Tools

    Get Citation

    Copy Citation Text

    Pengfei YE, Peng ZHANG, Hao YU, Shuang HE, Dongsheng TIAN, Yuanxin WANG, Shoufeng TONG. Deep Learning Aided Signal Detection Algorithm Experimental Research for Underwater Optical Communication[J]. Acta Photonica Sinica, 2024, 53(7): 0706001

    Download Citation

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

    Category: Fiber Optics and Optical Communications

    Received: Dec. 5, 2023

    Accepted: Jan. 16, 2024

    Published Online: Aug. 12, 2024

    The Author Email: ZHANG Peng (zhangpeng@cust.edu.cn)

    DOI:10.3788/gzxb20245307.0706001

    Topics