Deep Learning-Enabled Real-Time Recognition of Wireless Signals

In this demonstration, we perform a software defined radio (SDR) prototype of recognizing wireless signals in real-time. Convolutional Neural Network (CNN), a deep learning algorithm, is applied. Universal Software Radio Peripheral is used as RF front-end to transmit and receive over-the-air signals. MATLAB toolboxes are used to generate waveform for training, build learning models, train and test recognition and classification. Digital modulation types are recognized and classified in this demo. Meanwhile, we propose to demonstrate the classification of multi-carrier waveforms and multiple-in-multiple-out (MIMO) communications.