Publication Details:
Journal or Publication Title:
Armenian Journal of Physics=Ֆիզիկայի հայկական հանդես
Date of publication:
Volume:
Number:
ISSN:
Official URL:
Title:
Hybrid Bispectrum–Waterfall Feature Extraction with CS-DSB for RF Receiver
Creator:
Gasparyan, N. Y. ; Hakhumian, A. A. ; Sivolenko, E. R.
Subject:
Uncontrolled Keywords:
Specific Emitter Identification (SEI) ; Bispectrum ; Waterfall Features ; Time–Frequency Features
Coverage:
Abstract:
RF receiver identification requires robust feature extraction to distinguish subtle hardware-induced characteristics. Conventional methods based on higher-order spectra or time–frequency features often degrade under low SNR and multipath conditions. This paper proposes a hybrid bispectrum–waterfall feature extraction framework with CS-DSB (Carrier-Suppress Double Sideband) for RF receivers. The bispectrum highlights nonlinear phase coupling unique to receiver hardware, while waterfall features capture spectral and temporal variations. To improve efficiency, CS-DSB reduces data dimensionality while preserving discriminative information. A fusion scheme integrates both feature domains, followed by classification using a supervised learning model. Experimental results demonstrate that the proposed method significantly outperforms bispectrum-only, waterfall-only, and conventional approaches, achieving higher identification accuracy under noisy and bandwidth-limited scenarios. The findings show that combining bispectral, waterfall, and CS-DSB processing enhances robustness and enables efficient RF receiver fingerprinting.