Individual Identification Technology of Communication Radiation Sources Based on Deep Learning

Abstract

With the rapid development of wireless communication, the types of radiation sources are numerous and complex, and the signals are diversified. Traditional methods of identifying individual radiation sources may no longer meet the needs of society. The identification of individual wireless radiation sources is of great significance for ensuring the security of communication systems and improving the ability of military communication reconnaissance and countermeasures, but most of them use traditional identification methods. This article introduces deep learning as a classification method. Since there is no suitable public data set, we use 6 USRP devices of the same model, combined with LabVIEW software, in a laboratory environment to collect the IQ signals emitted by 5 individual radiation sources, perform preprocessing, and then invest in neural network training. The collected data can be used as a public data set for individual identification in the future. Using a variety of neural network structures and adjusting the parameters, we have obtained a more satisfactory classification effect.