Algorithm Research Of Multi-Array Passive Localization Based On M-Capon

Passive localization sensors have better ability to detect low altitude target and better survival in a hostile environment.In recent years,it has been studied by researcher and engineer.In a multi-array multi-target passive localization problem,the traditional two-step positioning methods need have complicated data association and cause the information loss by the intermediate parameter estimation.The direct position determination(DPD)methods have better positioning accuracy and stability under low signal-to-noise ratio by utilizing the original signal sampling for positioning compared with the traditional two-step positioning methods.However,the existing traditional DPD methods,such as ESPRIT and MUSIC method,require an exact number of targets in advance.In this thesis,the multi-array multi-target passive location has been studied under the condition of unknown number of targets.The main contents of this paper are given as follows.1.The traditional position methods including DOA,TOA,MUSIC and ESPRIT are introduced for the multi-array target passive localization and the corresponding Cramer-Rao lower bound(CRLB)is given.Then these methods are compared in the simulation.The location error is large of DOA in low SNR(Signal to Noise Ratio),while the MUSIC and ESPRIT direct position algorithms can obtain better location accuracy.2.A direct position determination algorithm based on m-Capon is proposed for the direct position problem under the condition of unknown target number.Its search spectral function is obtained by high-order inverse approximation of the autocorrelation covariance matrix of the multi-array received signal similar to MUSIC algorithm,then the target localization is achieved by searching in the diction space.The positioning error of this algorithm is analyzed with different parameters including SNR,snapshot number,sensor array element number and values of m in single-target and multi-targets scenarios.Furthermore the performance of the m-Capon direct position algorithm is compared with MUSIC,Capon algorithm.The simulation results shows that the positioning error of the proposed algorithm decreases with the increase of SNR,sampling snapshot number,array element number and m value.In the single-target scenario,the positioning performance of MUSIC,Capon and m-Capon direct location algorithms can reach the CRLB at the SNR of about-10 d B.In multi-target scenario,the positioning performance of m-Capon algorithm reaches the CRLB when the SNR is about-10 d B.3.A 3D direct position determination algorithm based on m-Capon is proposed to solve passive location in 3D scene.The algorithm increases the position accuracy of height estimation of target.The positioning error of this algorithm is analyzed with different parameters including SNR,target number and location array plane.The simulation results show that the positioning error of the proposed algorithm decreases with the increase of SNR.In the single-target scenario,the positioning performance of m-Capon direct location algorithms can reach the CRLB when the signal model is uniform array and SNR is about-10 d B,while the positioning performance of m-Capon direct location algorithms can reach the CRLB when the signal model is uniform cross-shaped array and SNR is about-5d B.In the multiple-targets scenario,there will be spikes and off-peak problems caused by the cross-intensive part of the space due to the increase of the number of targets under the uniform surface array and uniform cross-array signal model.These make the performance of the m-Capon algorithm decreases.