High Power Laser and Particle Beams, Volume. 35, Issue 6, 069001(2023)

Intelligent detection algorithm of broadband communication signal based on spectral decomposition

Dong Yi1, Ruipeng Ma2、*, Tao Hu1, Kaixin Cheng1, Di Wu1, Zhifu Tian1, and Yanyun Wang1
Author Affiliations
  • 1School of Data and Target Engineering, University of Information Engineering, Zhengzhou 450001 China
  • 2School of Cyberspace Security, Zhengzhou University, Zhengzhou 450002 China
  • show less
    Figures & Tables(13)
    Signal time spectrum diagram
    Structure of intelligent detection algorithm for large bandwidth time-width communication signal based on spectral decomposition
    Original image is directly fed into the target detection network
    Original image is decomposed and input into the target detection network
    Relative distribution of narrowband signal and spectral decomposition window
    YOLOx algorithm detection process
    The same IOU border overlapping situation contrast
    Diagram of prediction box fusion process
    Performance comparison of detection algorithms
    Performance comparison of intelligent detection algorithms under preclassification of signal targets
    Comparison of detection performance of the proposed algorithm under CIOU loss and IOU loss
    • Table 1. Results of experiments

      View table
      View in Article

      Table 1. Results of experiments

      parametervaluesparametervalues
      bandwidth range0~3200 kHzsignal typeswideband real signal
      sampling frequency6400 kHztime range2000 ms
      time-frequency spectrum frequency resolution1 kHztime-frequency spectrum time resolution1 ms
      size of time-frequency spectrum3200×3999the number of narrowband signals in a wideband signal40~70
      target signal frequency range5~3115 kHztarget signal bandwidth range5~100 kHz
      target signal modulation pattern2FSK, 4FSK, 8FSK, MPSKwhether there is a burst signalyes
      burst signal burst interval10~1500 msburst signal duration10~1500 ms
      parametervalues
      channel environmentRayleigh fading channel + non-stationary undulation noise + α-stabilized noise
      number of target signals in training set800 time-frequency spectrums, a total of 40 854 signals, 75% for training, 25% for validation
      number of target signals in test set0~20 dB stepped at 2 dB, and 100 time-frequency spectrums under each SNR, with a total of 61 373 signals
    • Table 2. Detection time of monad spectra detected by the network

      View table
      View in Article

      Table 2. Detection time of monad spectra detected by the network

      detection algorithmsingle sub-spectrum detection time/ms
      proposed algorithm45.0
      spectrogram decomposition combined with Ref. [16] 50.1
    Tools

    Get Citation

    Copy Citation Text

    Dong Yi, Ruipeng Ma, Tao Hu, Kaixin Cheng, Di Wu, Zhifu Tian, Yanyun Wang. Intelligent detection algorithm of broadband communication signal based on spectral decomposition[J]. High Power Laser and Particle Beams, 2023, 35(6): 069001

    Download Citation

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

    Category: Advanced Interdisciplinary Science

    Received: Jan. 11, 2023

    Accepted: Mar. 2, 2023

    Published Online: Jul. 10, 2023

    The Author Email: Ruipeng Ma (13164351610@163.com)

    DOI:10.11884/HPLPB202335.230024

    Topics