Experimental Validation of Spectral Sensing Models for Identification of Signals by Means of Artificial Intelligence

The paper is devoted to experimental validation of spectral sensing models for determining the information on the structure of a target signal by a cognitive radio receiver based on a neural network approach. The operating procedure of the LTE signal capture and marking models is described when scanning the radio airwaves using the hardware of a software defined radio board and software tools of the MATLAB environment. Deep learning models of a neural network of semantic segmentation of spectrogram images are used to identify the LTE signals.