Research On Polarization Signal Processing And Application Of Conformal Array

The conformal array is compatible with the aerodynamic design,leading to less friction,lower RCS and better stealth.In signal processing,it benefits wider coverage and surveil-lance for the configuration extends array aperture to the maximum extent.On the other hand,antennas fit into the curvature and orient distinctly,which yield individual antenna patterns and make the array polarization sensitive.The integrated information in space-time-polarization domain facilitates the performance of target detection and ground clut-ter suppression,which relies on utilizing the differences in space-time-polarization domain between the target echo and clutter,but in the challenge of parameter estimation in multi-dimension,calibration of steering error in the vector sensor array,non-homogeneous and non-stationary ground clutter suppresion.To deal with them,we propose fast parameter estimation algorithms,a calibration algorithm of antenna position error and one of mutual coupling,and a single data set STAP based on sparse recovery.The main contents are sum-marized as follows:1.For the curve configuration and diverse polarization responses present huge computa-tional burden on parameter estimation and customary fast algorithms designed for plane arrays fail due to the mismatched model,a fast DOA and polarization estimation algorithm based on the phase mode excitation theory(TID-DESPRIT)and one based on the Hough transform(PST-Hough)are proposed.TID-DESPRIT transforms the received data from the element domain to the phase mode(PM)domain,in order to use its mathematic structure and yield a closed form solution of parameter.After building the model of received data in the PM domain and trying to separate DOA and polarization,DOA is obtained via solving a quadratic eigenvalue problem built by the recursive relation of Bessel function and expo-nent item,and polarization is solved through the analytical form of PM model.PST-Hough solves parameters with the explicit analytical form of phases of the covariance matrix,and improves its accuracy via iterations.DOA is obtained via Hough transform using the co-variance matrix phases on the basis of a devised systematic scheme of ambiguity resolving.Subsequently,polarization is solved in a Rank-MUSIC like problem.Iteration handles with the effected accuracy due to the separation.Simulations validate their performance.2.The antenna position error not only changes the antenna spatial response but also po-larization response,while most customary calibration algorithms only focus on the former and their performance degrades when extended because of the dismatched signal model.To cope with that,a position error calibration algorithm in fractal domain based on calibrated signals is proposed.Firstly,the algorithm analyzes the specific effect of position error on spatial and polarization responses,and then separetes it into axial and height direction.The axial error is estimated firstly in a built equation with polarization responses substituted.The height error is calculated in a SCOP problem.Simulations validates the performance of the proposed algorithm.3.For the effects of mutual coupling(MC)on antenna patterns on horizontal(H)and verti-cal(V)planes are distinct and shall be modeled respectively,and the MC model matrix isn't of the Toeplitz structure which makes the related auto-calibration algorithm lose its efficacy,a MC calibration algorithm based on antenna pattern reconstruction with calibrated signals is proposed.The algorithm measures practical H and V patterns with calibrated signals whose parameters are carefully set,and then calculates the calibration matrix.The selection of parameters is analysed in terms of optimization and practical application.Simulations validates the performance of the proposed algorithm.4.The clutter of conformal array shows enhanced non-stationary,non-homogeneous char-acteristic and larger clutter rank compared to a plane array.Since then,phase compensa-tion algorithms obviously degrade in aligning the space-time spectrum of nearby range cell(NRC)and cell under test(CUT)for the configuration of conformal array is 3-dimentional.Besides,there are insufficient samples that are homogeneous and identically distributed like the CUT clutter.It concludes that the clutter distribution in the CUT are hard to be estimated from the NRC data.A single data set STAP based on sparse recovery(SR)for conformal array(SDSSR-CA)is proposed,where only the CUT echo is used to estimate the clutter covariance matrix(CCM)and sparse recovery handles the deficiency of CUT data.With the prediction of clutter distribution,several sectors on the AD plane are marked off according to clutter power,and the density of grid and the selection criterion of atoms in each sector are determined by the clutter distibution.The dictionary covering the clutter distribution is con-stituted of atoms selected in each sectors.Finally,the clutter data and CCM is reconstructed with the calculated overcomplete dictionary and AD profile.A STAP weight is obtained with the CCM involved.Simulations for real data validates the performance of SDSSR-CA.