Abstract
Accurate perceptual knowledge of atmospheric characteristics in the propagation path is of great significance for the design of communication systems. However, the atmosphere above the ocean is inhomogeneous, which brings challenges to accurate prediction of propagation loss. Moreover, the atmospheric refractive index distribution model calculated from atmospheric data requires at least two stations with near-sea meteorological data. In real maritime over-the-horizon communication or detection, only one station transmits and receives meteorological data, and the received meteorological data and propagation loss have large temporal noise. To address these issues, first, a denoising model based on the one-dimensional convolution autoencoder (1DCAE) is constructed to filter out the temporal noise of input environmental meteorological data and propagation loss. Second, to accurately predict the influence of environmental factors on the prediction of propagation loss, a deep-learning framework called short- and long- term parallel double-flow TrellisNets (SL-TrellisNets) is proposed to predict the loss. Finally, the extensive experiments demonstrated that the root mean square and mean absolute errors of the proposed 1DCAE are dramatically reduced. Our proposed SL-TrellisNets outperforms the other state-of-the-art techniques in terms of propagation loss prediction. In addition, we analyzed the impact of the environmental factors on the accuracy of over-the-horizon propagation loss prediction.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.