Reaserch On The Techniques Of Sepcific Emitter Identification For Pulse Signals

The specific emitter identification(SEI) processing theory and technique could hardly meet the requirements in demanding scenarios of the significant increase of the emitter system complexity, widespread application of the low probability of intercept(LPI) and the complex modulation. These realities, resulting in the loss of the effectiveness of the fingerprint features and the identification performance, caused inevitable huge difficulties to the existing SEI technologies. Many research results of the intuitive system evolutionary model in the field of nonlinear system modeling and nonlinear analysis show that the model is able to reveal the complex nonlinear characteristics of the system and its evolution. The intuitive system evolutionary is able to reveal the complex nonlinear characteristics of the system and its evolution regularity.Taking advantage of this model has important meaning to understanding and settling currently the plight of theory research and development and engineering practical problems in SEI field.Based on the SEI problem for pulse signals under complex electromagnetic environment as the rmajor esearch object, using the systemic Yoyos and stochastic differential geometry as analysis tool, we established the basic theory framework based on intrinsic geometrical characteristics of the emitter system state manifold for SEI. The validity criterion of fingerprint features and a fingerprint features extraction method based on the intrinsic instantaneous parameters(INIP-FEX) are proposed. These extracted features are actually intrinsic: changing coordinates for the signal will give a different description, but the same statistical properties. The INIP-FEX method can be directly applied in the SEI problems with simple radar pulses. Because of intrinsic fingerprint features integrated application of the intrinsic nature of the signal amplitude,phase, frequency parameters, therefore, the features will gain than the instantaneous frequency curve recognition of superior performance.To extend the intrinsic fingerprint features extraction method for complex modulation signals, and to increase the extracted feature accuracy of the emitter individual system state manifold geometric, we put forward the intrinsic fingerprint feature extraction method based on the diffusion maps(DIFF-IFEX) with the analysis on the flow of the vector field on individual emitter state manifold, and research results of the manifold learning field in recent years in this article. By using this method, we can estimate the signal intrinsic instantaneous parameters and extract intrinsic geometric features of individual emitter system by choosing an appropriate kernel for the diffusion process. Based on the intrinsic fingerprint features, which are extracted by DIFF-IFEX,we could have good identification performance in the linear frequency modulation,sinusoidal modulation signal environments.In the multipath signal transmission environment, the effectiveness of many existing fingerprint characteristics will be greatly reduced, leading to the SEI performance deteriorated badly. In order to solve this problem, this paper use the sparse property of multipath signal time delay domain, and introduce sparse reconstruction and variational Bayesian theories. Under the condition of known reference signals of the observation data of typical autocorrelation characteristic, we put forward a kind of multipath signal time delay estimation based on variable decibels Bayesian theory of emitter individual identification method(VB-TDE-SEI). SEI can be realized while multipath time delay is less than the width of the pulse by using VB-TDE-SEI. In the presence of unknown reference signal, we combine the intrinsic fingerprint feature extraction method based on the diffusion map and variational Bayesian theory, and put forward variational Bayesian theory and diffusion maps based feature extraction(VB-DIFF-IFEX) algorithm is proposed. The intrinsic features can be extracted from multipath signals observed data set, and applied to SEI. Simulation results show that the method could obtains a good SEI performance in conditions of unknown reference signal multipath observation.Because of the adverse impact of signal digital acquisition quantization errors,signal intrinsic instantaneous linearization approximation errors and the differences between different individual receivers, the performance based on the same fingerprint feature library maybe appears inconsistent in the application between two different receivers.In this visual system evolution model based on the framework of research,draw lessons from foreign receiver emitter individual identification field calibration technology. By using iterative algorithm and sampling reconstruction technology, we can compensate the receiver sampling frequency errors and channel power disturbance errors, and eliminate negative effects of individual differences in different receiver systems, that made a meaningful exploration of the engineering practice for SEI.