Research On Key Techniques In External Illuminator Passive Localization

External illuminator passive localization is a way of localization and tracking which exploits non-cooperative illuminators of opportunity as its transmitters, and deals with the direct wave received from the illuminator and echo wave reflected by the target through coherent processing. Then the localization and tracking process can be realized by estimating the related parameter information, such as time of arrival(TOA) and Doppler frequency from the echo wave of the target. Unlike the radar and traditional target illuminator that confront the huge challenges and threats recently, passive localization based on external illuminator has broader spatial coverage space, strong anti-interference ability, good sheltered performance and a better ability to detect the stealth target, which can improve the weapon’s capability of surviving and battling in the environment of electronic warfare. Therefore, it causes widespread attention and deeply research around the world. This paper mainly focuses on the time delay estimation, frequency estimation and positioning calculation in the passive localization based on external illuminator in order to meet the requirements of a specified project. The main research work and contributions are outlined as follows:The relationship among direction of arrival(DOA), time delay estimation and frequency estimation are investigated. On the basis of deeply analyzing these three parameter estimation models, the uniformity of these models has been proved assuming that the uniform line array(ULA) is employed which illustrates that their estimation methods can be applied generally. Then the time delay estimation and frequency estimation based on subspace theory and sparse decomposition theory are studied. What’s more, their corresponding Cramér-Rao lower bounds(CRLB) are derived as the criteria to judge the performance of algorithms.An improved signal subspace scaled multiple signal classification(SSMUSIC) is proposed to solve the problem of performance degradation of the traditional subspace methods at the lower signal noise ratio(SNR). It enhances the spectrum’s resolution and accuracy by smoothing the correlation data and weighing orthogonal subspace. A time delay estimation method based on the weighted subspace fitting is proposed, which uses model transforming to make the traditional spatial spectrum estimation method applied to time delay estimation, and combines the fast algorithm of modified variable projection(MVP) to solve the optimization problem. Simulations show this algorithm has a better spectrum’s resolution and accuracy at low SNR. Moreover, an improved frequency estimation algorithm based on the maximum eigenvector is proposed, which uses the maximum eigenvector as the sparse decomposition vector, then the problem is transformed to a second order cone(SOC) problem and thus the estimation is obtained by utilizing optimizing algorithm. Simulations show the improved algorithm has high resolution ability.The study is made with regard to multidimensional scaling(MDS) positioning algorithm. On the basis of analyzing the present real multidimensional scaling positioning algorithm, a unified framework of complex MDS is proposed in this paper. The algorithm constructs the complex scalar product matrix with complex coordination, and then gets the noise subspace by singular value decomposition(SVD). The corresponding complex version of the traditional MDS algorithms can be established with different weighted vectors, which improves the positioning accuracy effectively. The complex coordinates extend the dimension of noise subspace and strengthen the constraints between observation stations and targets. Finally, the simulations compare the performance of real MDS algorithms and complex MDS algorithms, and thus verify the effectiveness of the algorithms.The research is made in regard to the direct position determination(DPD). The estimation of target’s coordinates can be treated as the estimation of location parameter. DPD algorithm can achieve the target’s coordinates directly from the received signals, which avoids the mismatch between target’s practical coordinates and parameter estimation of the traditional two-step positioning algorithm, and the problem of measurement errors being amplified. Therefore, the superior performance can be realized. On the basis of the studies on the existing DPD algorithm, this paper proposes an improved algorithm by combining the Gaussian maximum likelihood model. Firstly, it constructs Gaussian maximum likelihood estimator after transforming the data to the frequency domain. Then we take the procedure of eigenvalue decomposition for the Hermitain matrix which contains the target’s coordinates. The problem thus can be transformed to calculate the maximum eigenvalue. Finally, the target’s coordinates can be estimated through the search of geo-grids. The improved algorithm utilizes the information of each receiver adequately, which enhances the accuracy of positioning and attains the Cramér-Rao lower bound. The simulations verify the validity of the improved algorithm and the cases of multi-targets are also analyzed.