Here are some resources on Software Defined Radios (SDR) and using deep learning with them.

SDR is used in the Search for ExtraTerrestrial Intelligence (SETI) for its flexibility. Here is a good example of recognizing  patterns on a spectrogram (frequency/time plot) that just made the news:

Distant galaxy sends out 15 high-energy radio bursts

Distant galaxy sends out 15 high-energy radio bursts

Tinkerers found out that one can convert a cheap (like $25) digital TV dongle using the RTL2382U chip into a software defined radio and get them to work with gqrx (linux) or SDR#. Thanks to the Berkeley SETI website for tipping me off to this.

GQRX Preview

https://www.youtube.com/watch?v=PcBSNWiJjHk

RTL-SDR Dongle and GQRX on Mac OS X

https://www.youtube.com/watch?v=bSAa2aOXpCc

History and Discovery of RTLSDR
http://rtlsdr.org/#history_and_discovery_of_rtlsdr

Wikipedia page on SDR
https://en.wikipedia.org/wiki/Software-defined_radio

RadioML

https://radioml.org/ & https://github.com/radioML/

https://oshearesearch.com/

RadioML docker image

https://github.com/radioML/dockerRML

 

Guess what? This space of pattern recognition in frequency/time