RFFsNet-SEI: A multidimensional balanced-RFFs deep neural network framework for specific emitter identification

Existing specific emitter identification (SEI) methods based on hand-crafted features have drawbacks of losing feature information and involving multiple processing stages, which reduce the accuracy emitters complicate procedures identification. In this paper, we propose a deep SEI approach via multidimensional extraction for radio frequency fingerprints (RFFs), namely, RFFsNet-SEI. Particularly, extract physical RFFs from received signal by virtue variational mode decomposition (VMD) Hilbert transform (HT). The I-Q data are formed into balanced-RFFs, then used to train As introducing model-aided neural network, hybrid-driven scheme including is constructed. It improves interpretability Meanwhile, since RFFsNet-SEI identifies individual raw in end-to-end, it accelerates implementation simplifies Moreover, as temporal spectral both extracted RFFsNet-SEI, improved. Finally, compare with counterparts terms accuracy, computational complexity, prediction speed. Experimental results illustrate that proposed method outperforms basis simulation dataset real collected anechoic chamber.