• DocumentCode
    357803
  • Title

    Using artificial neural network approach to predict rain attenuation on Earth-space path

  • Author

    Hongwei Yang ; Chen He ; Wentao Song ; Hongwen Zhu

  • Author_Institution
    Dept. of Electron. Eng., Shanghai Jiaotong Univ., China
  • Volume
    2
  • fYear
    2000
  • fDate
    16-21 July 2000
  • Firstpage
    1058
  • Abstract
    A rain attenuation model based on artificial neural network is proposed in this paper without making many assumptions as traditional methods and therefore improves the prediction accuracy. Based on an analysis of various factors affecting rain attenuation, a rain attenuation model with artificial neural network is developed after training and verifying many different neural network topologies. The prediction results of the proposed model is also compared with that of the CCIR model. The results show that applying the artificial neural network to predict rain attenuation of a high frequency wave is a good approach and decreases the mean prediction error by 0.59 dB and the RMS error by 0.69 dB. The paper shows that this model is a new and effective way to predict rain attenuation with an artificial neural network.
  • Keywords
    HF radio propagation; backpropagation; electromagnetic wave absorption; neural nets; rain; space communication links; telecommunication computing; tropospheric electromagnetic wave propagation; ANN; CCIR model; Earth-space path; RMS error; artificial neural network; backpropagation neural network; high frequency wave; mean prediction error reduction; neural network topologies; prediction accuracy; radiowave propagation; rain attenuation model; rain attenuation prediction; training; training data; Accuracy; Artificial neural networks; Atmospheric modeling; Attenuation measurement; Communication systems; Frequency; Helium; Meteorology; Predictive models; Rain;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Antennas and Propagation Society International Symposium, 2000. IEEE
  • Conference_Location
    Salt Lake City, UT, USA
  • Print_ISBN
    0-7803-6369-8
  • Type

    conf

  • DOI
    10.1109/APS.2000.875404
  • Filename
    875404