Research On Imaging Localization Method Of Radiation Sources Based On Synthetic Aperture System

A passive localization system locates a radiation source by the passively received signal.Since it does not emit electromagnetic waves,it has the advantages of high concealment and long-acting distance and has wide applications in the fields of long-range early warning,ground air defense,and electromagnetic environment monitoring.Moving-single-station radiation source localization is highly mobile and can be flexibly deployed.Radiation source localization with a single moving platform has a promising application in military and civilian fields.To meet the demand for high-resolution and high-precision localization,the existing single-station localization system needs to carry as large an array of antennas as possible,but the weight-carrying capacity of the platform determines that the aperture of the antennas is limited.Current methods for passive localization of radiation sources from moving single stations mainly focus on algorithms of spectral estimation.The existing algorithms obtain the radiation source location by the amplitude accumulation of the spectral estimation results from different locations.This is a non-coherent superposition approach,which affects the resolution of localization and causes large errors under low signal-to-noise ratio(SNR)conditions.To meet the demand for high-resolution and high-precision localization of radiation sources,this dissertation introduces the idea of imaging in synthetic aperture radar and the technique of coherent processing and carries out the research of imaging localization based on synthetic aperture technology.The working system,mathematical principles,localization methods,and practical applications are carefully studied.Starting from the principle of maximum likelihood estimation,the dissertation researched the problems of multiple radiation sources,high-efficiency localization,multi-angle localization,and ultra-high resolution in syntheticaperture-based localization.The main work is as follows:1.The geometric model of synthetic aperture localization,the working system,and the cost function based on the maximum likelihood estimation are established,which lay the foundation for the subsequent research of localization methods and the analysis of localization performance.First,the geometric model of localization is constructed based on the variation of the distance between the receiving platform and the radiation source during the motion.The proposed new localization system adopts a two-dimensional data sampling mode,and the two-dimensional data matrix is described by range time and azimuth time.By the coherent accumulation of two-dimensional data,a cost function based on the maximum likelihood estimation is constructed and the mathematical principle of localization is clarified.In addition,the advantages of coherent processing under low SNR conditions are demonstrated by output SNR analysis,and the Cramer-Rao Lower Bound(CRLB)of this localization system is analyzed.Finally,an algorithm based on maximum likelihood estimation is proposed to complete the high-resolution high-precision localization of radiation sources.2.A passive signal focusing algorithm for synthetic aperture localization of multiple radiation sources is proposed to achieve simultaneous imaging and localization of multiple radiation sources.The algorithm of the maximum likelihood estimation class contains a twodimensional search process,which takes a lot of time in the face of large scenes and multiple radiation sources localization.To address this problem,a passive signal-focusing algorithm for multi-target synthetic aperture localization is proposed.The range focusing is achieved by eliminating the baseband modulation of multiple radiation source signals through the higher-order spectrum of the signals,and the matching filtering is used to focus the multiple radiation source signals in the azimuth domain.According to the focusing result,multiple radiation source positions can be obtained.To deal with the problem of cross terms in the high order spectrum,a CLEAN-based multi-target high separation localization algorithm is proposed.A matched filtering method is used to focus the signal with low modulation order to obtain the range position,azimuth position,and frequency of the radiation source.Then the defocused signal is filtered.Finally,the inverse function of the matched filter is used to recover the signal.The process is repeated according to the modulation order to obtain high separation localization results.3.A Keystone-Deramp localization method based on time-frequency distribution is proposed to achieve efficient localization in the case of large scenes with high resolution.First,we reconstruct the quadratic signal matrix in the delay-azimuth time domain by range focusing method and autocorrelation processing of the azimuth signal.The coupling of the delay and azimuth time in the quadratic signal matrix causes the signal to be incompletely focused.The Keystone transform is used to eliminate the two-dimensional time coupling,and the Deramp processing is used to eliminate the localization aliasing in the range domain.Finally,the signal in the two-dimensional domain is converted to the two-dimensional spatial domain by the scaled fast Fourier transform to complete the localization of the radiation source.4.A spaceborne synthetic aperture localization method for long-time data is proposed to reduce the influence of the carrier frequency estimation error on localization accuracy.First,the high-order slant range model from the receiving platform to the radiation source is established.Then,the azimuth matching filter is constructed by the de-chirp method and the high-order term modulation rate of the received signal is estimated by the method of azimuthal focusing and the minimum entropy method.Finally,the grid is divided under the geographic coordinate system and the cost function is constructed to locate the radiation source.Moreover,a localization performance analysis method based on higher-order term orthogonality is proposed,which clarifies the localization capability of the method for radiation sources at different locations.5.A spaceborne ultra-high resolution synthetic aperture localization algorithm for ultra-longtime data is proposed to achieve ultra-high resolution radiation source localization.First,the unbiased estimator of the radiation source carrier frequency based on the sub-aperture chirp rate fitting is proposed.The chirp rates of sub-apertures are estimated by the method of minimum entropy and the fitted result of the estimated value determines the zero Doppler moment of the radiated source signal so that an unbiased estimate of the carrier frequency can be obtained.Then,the Doppler spread is eliminated by compensating the instantaneous Doppler of the received signal based on the fitted chirp rate estimates.The azimuthal fullaperture signal is extracted by range-focusing processing and then divided into sub-apertures.The problem of the insufficient sampling rate of the sub-aperture is solved by the spectral compression method.Finally,based on the localized two-dimensional resolution and beam pointing,a non-uniform sampling grid is divided to calculate the localized image of each sub-aperture.The coarse estimates of the radiation sources are obtained in a non-coherent summation manner.In the neighborhood of the coarse estimate,a fine sampling grid is divided and the initial phases of different sub-aperture images are compensated to complete the coherent summation of all sub-aperture images and obtain an accurate estimate of the location of the radiation source.