Study Of Symmetric Nonuniform Linear Array For Localization Of Mixed Near-and Far-Field Sources

Signal source localization technology is an important part of array signal processing and has been widely used in many fields of military and people’s livelihood.Compared with the common far or near field source localization,mixed field source localization is more complex,which requires the estimation of direction of arrival(DOA)and near field distance parameters,as well as the classification of near and far field sources.Most of the existing mixed source localization techniques are implemented based on symmetric array(SA).However,due to the design limitation of symmetry,the degree of freedom(DOF)and physical aperture of array are inhibited.In the SA structure,the inhibition of mutual coupling effect between elements still needs further study.In this paper,we study the design of symmetric nonuniform linear array(SNLA)configuration to increase the DOF of array and suppress the coupling effect.Combining with low complexity algorithm high precision location of mixed field sources can be achieved.The main work of this paper is as follows:Firstly,the unified form of the existing SAs are summarized,and the structure design of the even element part of the partial SAs are improved.Then their structural characteristics and mutual coupling performance are compared and analyzed.Based on the existing symmetric nested arrays,two kinds of SNLAs with high DOF,namely symmetric displaced composite nested array(SDCNA)and symmetric extended nested array(SENA),are improved and designed.Their optimal design schemes are derived.The SDCNA extends the central subarray into sparse uniform array,and carries out composite nesting on both sides.This design improves the continuous degree of freedom,and effectively inhibits the coupling effect of the array.The SENA improves the degree of freedom,and further expands the array aperture.Considering the mutual coupling effect between array elements,a symmetric displaced extended nested array(SDENA)with high DOF and low mutual coupling is designed.In this array,the sparse central subarray is symmetrically flipped and displaced,and then the interelement spacing of nested subarrays on both sides is enlarged.And the holes of the difference coarray(DCA)are supplemented by the extended part.The array effectively suppressed the mutual coupling effect by enlarging the interelement spacing.The coupling leakage of SDENA is lower than that of other arrays,and shows a decreasing trend with the increase of array number.The extended part of the array configuration enables the array to obtain higher continuous DOF,and more mixed field sources can be detected at the same time.The SAs are combined with the algorithm with low computational cost to realize high precision mixed source location.Firstly,the fourth order cumulant matrix excluding the distance parameters of near field sources is constructed by using the symmetry property of the array.And the DOAs of mixed sources are estimated by combining with the root-MUSIC algorithm under spatial smoothing.The estimated DOAs combined with the MUSIC algorithm are used to search the spectral peak of the distance of the near field source.Comparing the range estimates with the near field area,the far and near field sources can be classified.The mixed source two-dimensional parameter estimation is transformed into two one-dimensional searches,and combined with the root-MUSIC algorithm to avoid DOA spectrum search,which effectively reduces the amount of computation.In this paper,the simulation of mixed field source localization is carried out by combining the above algorithms with different SNLAs.The effects of signal to noise ratio,snapshot number and mutual coupling on the accuracy of mixed source localization are analyzed.The results show that in the high DOF arrays,the mixed source estimation performance of the SENA is better than other arrays without considering the mutual coupling effect,followed by the SDCNA.Under the mutual coupling condition,the mixed source localization effect of SDENA is better than other arrays,followed by SDCNA.