Reseach On The Signal Emitter Individual Identification Technology Of The Fingerprint

The radiation source fingerprint identification technology is signal processing in a new research direction, were analyzed and identified by radar signal propagation on the wide range, the technology can be widely used in military and civil fields. The domestic and foreign research mainly aims at the transient start-up signal, the feature extraction is often has its limitations, so this paper of different sources of same type and working mode, is studied through analysis of steady-state fine individual communication signals feature extraction technology. The main work of this paper includes the following 3 aspects:1 Studied the extraction technology of code rate and carrier frequency based feature, improved the phase differential carrier frequency estimation algorithm, using the phase difference method, do two times smoothing the noise phase increment, reduces the phase noise of carrier frequency estimation, and the method has small amount of computation, simple and easy to implement. Based on the wavelet transform and STFT time-frequency energy distribution improved estimation method solves the precision problem of code rate in non-cooperative communication environment. Theoretical analysis and simulation results show that, under the same conditions, the bit rate and frequency estimation precision is improved compared with the traditional estimation method, can meet the extraction of radio signal carrier frequency characteristics of the interception requirements.2 The signal fingerprint analysis and extraction of radiation source based on composition of individual spurious output. Extraction method of Hilbert transformation method, Teager-Kaiser method and orthogonal component reconstruction envelope gives, envelope modulation feature extraction and analysis method and the box dimension feature, dimension and Lempel-Ziv fractal complexity characteristics. And using the wavelet de-noising method to reduce the effect of noise on the intercepted signal, so as to improve the efficiency and accuracy of the empirical mode decomposition method, and then use Hilbert edge the intercepted signal spectrum symmetry features and Hilbert Huang transform time-frequency distribution image extracting fingerprint feature signal. Experiments show that different signal spectrum symmetry between feature vectors vary greatly, the same signal spectrum symmetry features vector difference between the smaller, better clustering results using this feature, application of radiation source signal in the identification and classification of different.3 Reseach on higher order spectrum of signal are studied fingerprint analysis and extraction. Based on four kinds of partial double spectrum, and the rectangular integral bispectrum signal fingerprint feature extraction method based on higher order spectrum, and by using Fisher identification function simulation verified the bispectrum analysis method in the performance of the signal feature extraction. Experiments show that the rectangular integral bispectrum than local bispectrum analysis method, in terms of individual to extract the signal characteristics of performance is better.