Research On MIMO Radar Imaging For Sparse Distributed Target

Multiple input multiple output(MIMO)radar can utilize the space-wide multiple transmitting and receiving arrays to significantly expand the spatial sampling capability of target's scattering information,improve the spatial resolution of radar system and re-alize the imaging observation of target.The coherent MIMO radar with compact spatial distributed arrays is a main kind of application form for MIMO radar.The research of coherent MIMO radar imaging technology has been paid more and more attention in the fields of target surveillance,safety detection,remote sensing and so on.The multi-channel spatial samplings of the coherent MIMO radar have partial correlation due to the use of compact element space distribution,which becomes the bottleneck of further improving the spatial resolution of radar system.In this paper,based on the characteris-tic of spatial spectrum filling for coherent MIMO radar and through introducing a sparse priori to the target,the following research work is carried out to solve the bottleneck problem:Firstly,the two-dimensional echo model of coherent MIMO radar under the far field condition is deduced.The characteristic of spatial spectrum and its influence on the imaging performance of traditional matched filtering method are analyzed based on the spatial spectrum theory.Then the MIMO radar simulation and experimental scenarios are constructed for the subsequent research contents.Secondly,we propose the method based on matched filtering and regularization to improve SNR of MIMO radar.Moreover,according to the correlation of coherent MIMO radar echo,a low rank matrix denoising algorithm combing with nonconvex constraint is proposed,which can further improve the echo SNR.Thirdly,a two-dimensional convolution signal model of point spread function(PSF)and target's scattering coefficient is constructed under the situation of uniform and com-pact spatial spectrum filling.Two-dimensional iterative deconvolution imaging meth-ods are proposed to eliminate the convolution effect of PSF.Next,two kinds of echo matrix rearrangement method are proposed for the situation of uniform spatial spec-trum with some rows and columns missing randomly due to reducing the number of arrays and transmitting frequency.The low rank matrix completion(MC)method with nonconvex constraint is proposed to fill the missing spatial spectrum based on the cor-relation between rearranged echo.Fourthly,under the situation of non-uniform spatial spectrum filling which can not be completed by MC,the method of deconvolution is not applicable.Thus,a two di-mensional sparse imaging method with nonconvex constraint is proposed based on the theory of compressed sensing(CS).The influence of model mismatch error caused by off-grid and phase error on the imaging performance is analyzed by simulation.Then,the coherence characteristic of the observation matrix is discussed from the imaging grids and echo sampling redundancy caused by the correlation of spatial spectrum.For the problem of observation matrix with lot of redundant data and bad coherence,The projection matrix optimization method is proposed to reduce the coherence of observa-tion matrix.