Repository for my learning examples, maybe used as pieces for other projects.
Simple DNN approach for a classifier of signal or FM signal, based on a raw iq samples value small sample.
Working on CNN approach for a classifier of signal or FM signal (stripped, minimal), based on a raw iq samples value small sample. This example is designed to recognize presence of signals in a 250mhz bandwidth sample. It takes 2048 samples of it to decide between four classes: Noise, FM, GSM and Carrier (simple carrier or peaks in the spectrum).
Files for capturing data. Actually gnu radio companion is used (SIGID-rtl-window-capture-base.GRC), but rtl-capture.py should be used but it is not actually working.
Not used yed, planned to input waterfall and/or sdft as input.