Abstract
Spectrograms are an essential time-frequency representation tool that has been used to address several important problems in wireless communication systems. However, most existing techniques based on the processing of radio spectrograms require a relatively high Signal-to-Noise Ratio (SNR), performing poorly at low/moderate SNR. In this context, this letter proposes an iterative radio spectrogram filtering method based on a novel pyramidal convolution kernel. The obtained results demonstrate that the proposed technique improves the recognisability of signal components in radio spectrograms. Two illustrative examples are provided to show how this method helps to extend noticeably the SNR operational range of techniques for wireless communications based on the processing of radio spectrograms.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
