A New Method of Specific Emitter Feature Extraction Based on IQ Imbalance
Peng Wang, Jinmimg Wang, Yuan Sun, Hongyu Zhang · 2021
A new method for fingerprint feature extraction of specific emitter is proposed, the method used the individual differences of gain mismatch and phase mismatch which based on the I/Q imbalance. Received signal is denoised, demodulated and filtered, and the I/Q information of the original signal is decomposed in this method. Respectively calculate the ratios of autocorrelation functions to the cross-correlation function of I/Q signals at 0, and two specific values only related to gain and phase mismatch parameters. The values are used as two dimensional characteristic parameters and the parameters are used in specific emitter identification. Compared with the method based on bispectrum and Hilbert-huang transform method, the identification accuracy of new method is improved by more than 5% and 10% under the condition of simulation experiment. Compared with the SNR estimation method based on eigenvalue decomposition, the identification accuracy of the proposed method is improved by 3% and identification effect is better under the condition of low SNR.