Research On Applications Of Chaotic Signal Processing In Specific Emitter Identification

Specific emitter identification(SEI)identifies the individual emitter of a given signal by using only the radio frequency(RF)fingerprints carried by the signal.This technique is often used to enhance the security of wireless network and it is also helpful in military applications.However,it is not an easy task to perform an effective SEI system,and it becomes even more difficult to extract effective RF fingerprints from received signals in these days since the number of communication emitters is growing quickly and communication environment is becoming more and more complex.In this dissertation,a received signal is considered as an output of a nonlinear dynamical system which is consisted of the transmitter,the transmission environment and the receiver.Based on this principle,this dissertation has presented some methods based on chaotic theory to address the SEI problem and all these methods are tested by using data from real applications.The dissertation deals with the fundamental chaotic theory as well as the applications of chaotic theory in SEI problem,the main contributions of this dissertation are as follows:1.A procedure of component reordering is presented by analyzing the incomplete reconstructed phase-space.The procedure can transfer the information captured by the incomplete reconstructed phase-space into the reordered component and subsequently reveal some basic properties of chaotic series.Moreover,the incomplete reconstruction can avoid some challenges caused by the delay-coordinate reconstruction.2.Chaotic signal detection based on the incomplete reconstructed phase-space is presented.In order to distinguish between chaotic series and additive white Gaussian noise in the case of small amount of data and low signal-to-noise Ratio(SNR),characteristics of different series are first analyzed in the incomplete reconstructed phase-space.Then a statistic named segmented mean variance(SMV)is presented to quickly distinguish between chaotic series and Gaussian white noise based on the fact that the SMV of Gaussian white noise obeys F-distribution.In order to distinguish among chaotic series,correlated and uncorrelated stochastic series,a method based on component reordering and the visibility graph(VG)is introduced.Time series and their reordered components show diverse characteristics in their degree distributions,then the Kullback–Leible divergence is used to quantify the difference between these two degree distributions and to distinguish among chaotic systems,correlated and uncorrelated stochastic processes.3.The delay of time-delay chaotic system is estimated by using a method based on the incomplete reconstructed phase-space.The local clustering phenomenon of the orbits in incomplete reconstructed phase-space is analyzed.The relationship between the embedding delay of incomplete reconstruction and the delay of chaotic system is investigated.Then,the intrinsic delay of chaotic system is quickly extracted from the reordered component by using the SMV in the case of small amount of data and low SNR.4.SEI methods based on fundamental chaotic features are investigated.The first method uses the information captured by the incomplete reconstructed phase-space.In order to discriminate different emitters,the SMV calculated from signal's reordered component is utilized as the RF fingerprint.Then some real data sets are used to test the performance of the SMV-based method.Moreover,in order to verify the assumption that the transmitter,the environment and the receiver can be can regarded as a nonlinear dynamical system,the effect of multipath on the method is tested by using the data from a real application.The second SEI method is based on the nature measure(NM)of chaotic signals.The method shows that nature measures of the one-dimensional components of higher dimensional systems exist and thus they can be used as RF fingerprints to identify different emitters.The NM-based method is also evaluated by using data from real application.5.SEI methods based on the nonlinear characteristics captured by the VG are investigated.In order to identify the emitters produced by the same manufacturer with high recognition rate,three approaches based on the VG and horizontal visibility graph(HVG)are presented.In the first two approaches,the VG entropy or the HVG entropy of the corresponding degree distribution is used as the RF fingerprint to identify different emitters.In the third method,in order to improve the performance of the HVG-based features,both the HVG entropy and the Fisher's information measure(FIM)are extracted from the corresponding degree distribution to form a feature vector and subsequently to identify different emitters.These methods are verified by some real data sets.Moreover,effects of multipath on the proposed methods are investigated.