DOA Estimation Based On Expanding The Array Aperture

Array aperture is one of the most important parameters in array signal processing.Expanding the effective array aperture is a significant method to improve the estimation accuracy in spatial signal processing.In this thesis,direction-of-arrival(DOA)estimation methods based on expanding the array aperture are investigated,and three DOA estimation algorithms are proposed,including Rec-ESPRIT algorithm for noncircular signals,DOA estimation method for quasi-stationary signals with Huang-nested array,and sparse representation based DOA estimation algorithm with a modified nested linear array(MNLA).Rec-ESPRIT algorithm is proposed to deal with the problem of DOA estimation of noncircular signals with a partly calibrated uniform linear array.Taking advantage of the non-circularity of the signals,an extended data model corresponding to a virtual(extended)array is presented.On this basis and given the signal subspace matrix associated with the virtual array,a linear equation is derived to determine the unknown gains and phases.Then,the DOAs are found through the eigenvalue decomposition of a matrix related to the signal subspace matrix and array gains and phases.Since spatial spectrum search is not required,the proposed method is c omputationally efficient.Moreover,it is able to handle at most 2M-3 noncircular signals with a partly calibrated ULA of M elements,even though only two neighboring elements of the array are calibrated.Huang-nested array is a new array configuration based on nested array,and it is proposed in order to increase both the virtual array apertures(VAAs)and degrees of freedom(DOFs).Huang-nested array configuration is composed of two independent ULAs plus an additional sensor,and all sensors are arranged in linear.Assuming the total number of physical sensors is L and the inter-element spacing of the first ULA is d1,it will be shown that the DOFs and VAAs of the proposed array configuration are,respectively,2L-6 and nearly(L2/2-L-2)d1 larger than those of the original two-level nested array.Such increase in both DOF and VAAs would lead to enhanced resolution,which enables us to estimate the signal parameters such as DOA more accurately.Recently,Yang et al.have proposed a modified nested linear array(MNLA),which has a longer consecutive virtual sensors in the resulting difference co-array,compared with that of the traditional nested linear array(NLA).In this paper,we investigate the problem of DOA estimation with MNLA base on sparse representation.We shall show that the maximal level of detectable sources that can be achieved by a MNLA with L physical sensors is almost (L2-2)/4+L-1.Moreover,it is shown that better performance of DOA estimation can be obtained through sparse representation than spatial smoothing.