Modulation recognition with GNU radio, keras, and HackRF

Abstract

Recognition and classification of the modulation format of received signals in real time is a challenging but necessary task for the intelligence community. Without prior knowledge of the received data, such as power, frequency, and phase, coupled with real-world concerns such as signal degradation and interference, the task of reconstructing the sent information is made even more complex. Our approach for classifying modulation types of live over-the-air signals, a precursor step to demodulation, involves using software defined radio plus two simple and inexpensive transceivers to train a series of neural networks over a set of three data types.