APPLYING ARTIFICIAL NEURAL NETWORKS IN THE TASKS OF DETERMINING THE TYPE OF SIGNAL MODULATION

This paper considers the application of convolutional neural networks for automatic classification of modulation in wireless communication systems. The importance of the task of automatic recognition of modulation type is shown. Traditional methods of modulation recognition are considered, and the limitations of traditional methods are shown. The study examines the effectiveness of deep learning methods, in particular CNN, to recognize modulation types such as BPSK, QPSK, 8-PSK, 16-QAM, 64-QAM and others. The proposed model is trained on synthetic datasets with different levels of signal-to-noise ratio (SNR). The architecture of the model is described. The results of verification of the accuracy of recognizing the type of modulation in conditions of low SNR are shown. The perspicuity of using artificial neural networks to solve the problem of recognizing the type of modulation is shown.