A Specific Emitter Identification Method Based on RF-DNA and XGBoost

Abstract

Specific emitter identification (SEI) is a promising research direction in artificial intelligence and Internet of Things. Aiming at the fingerprint features extraction and the identification algorithm selection for SEI, a novel method based on RF-DNA feature set and extreme gradient boosting (XGBoost) algorithm is proposed in this paper. Firstly, considering the advantages of RF-DNA in characterizing the fluctuation degree of instantaneous sequences, a RF-DNA feature set is constructed based on statistical features extracted from instantaneous frequency, instantaneous phase and instantaneous amplitude of signals. Then, the XGBoost algorithm is used to perform feature learning on the structured RF-DNA data set. Finally, three civil communication emitters of the same model are used as the identification objects to verify the performance of the identification method. Experimental results show that the RFDNA feature set exhibits satisfactory feature expression performance, and the XGBoost algorithm shows favorable feature learning properties.