Specific emitter identification using decomposed hierarchical feature extraction methods

Abstract

Specific Emitter Identification(SEI) has drawn wide attention in the field of reconnaissance and communication. Stray features which can be used as the emitter “fingerprints” usually hide deeply in the received signal. Since traditional feature extraction methods are executed on the original level of received signal, which may lose the crucial hidden information for SEI. Finding out those unique characters to improve the SEI performance become an emergency. In this paper, we introduce the intrinsic time-scale decomposition (ITD) method into our feature extraction process. Stray features such as energy distribution information, fractal features and bispectrum features are extracted on the decomposed levels of signal. Support Vector Machine(SVM) is also applied to make the classification. Experiments on real data verified that the proposed method has a remarkable performance compared with other methods.