Abstract
Predicting the channel quality for an underwater acoustic communication link is not a straightforward task. Previous approaches have focused on either physical observations of weather or engineered signal features, some of which require substantial processing to obtain. This work applies a convolutional neural network to the channel impulse responses, allowing the network to learn the features that are useful in predicting the channel quality. Results obtained are comparable or better than conventional supervised learning models, depending on the dataset. The universality of the learned features is also demonstrated by strong prediction performance when transferring from a more complex underwater acoustic channel to a simpler one.
Highlights
Underwater acoustic (UWA) propagation is a complex probabilistic process dependent on many time-varying factors
In our previous work [8], traditional engineered features were derived from the full orthogonal frequency-division multiplexing (OFDM) waveform and were used with conventional regression methods to predict the communication performance
For the purposes of this paper, traditional features refers to signal statistics derived from the OFDM waveform, such as signal-to-noise ratio (SNR) and delay spread
Summary
Underwater acoustic (UWA) propagation is a complex probabilistic process dependent on many time-varying factors. For the general communication system, it is often desirable to establish a relationship between channel characteristics and the receiver decoding performance; the complex UWA environment often makes this a challenging task. Reliable prediction of the UWA communication performance is an enabling technology for many useful methods and systems, including adaptive networking and adaptive modulation strategies [1,2], for both point-to-point communications and underwater sensor networks; please refer to [3,4] for corresponding system characteristics and challenges. Previous works have focused largely on utilizing meteorological data [5] or signal statistics as input features [6] These features are either observed directly through additional sensors, such as the meteorological data, or are engineered features estimated from signal statistics. A boosted regression tree was used in [2] to estimate the BER based on channel parameters such as the signal-to-noise ratio (SNR) and the delay spread, in order to apply an adaptive modulation scheme
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.