The Very First Attempt to RF Signals Detection in Wide RF Spectrum Using Deep Learning

The article presents the initial stage of research on the application of deep learning machine to the detection of radio signals in the broadband radio spectrum. A measuring station based on a software defined radio and a PC with GNU Radio and PyTorch software was put together. During the research, positive results were noted with the use of a trained with low-randomized data, Mask R-CNN network [1] for the detection of radio signals in the radio spectrum.