The Research Of The Beamforming And DOA Estimation Algorithms For Comformal Phased Array

The array signal processing,an important branch of signal processing, mainlyincludes adaptive beamforming and DOA estimation techniques. Adaptivebeamforming and DOA estimation techniques, which can improve the performance ofthe array antenna, have broad application prospects in military and civilian applications.The foundation of the array signal processing is the array configuration. Due to theconsideration of aerodynamics or hydraulics, the comformal array is able to furtherimprove the performance of the system. The uniform circular array (UCA) and thespherical array, two typical configurations of the confomal array, are able to provide360o of coverage in the azimuth plane and have uniform performance regardless ofangle of arrival. Moreover, they can provide the elevation angle information. Therefore,compared with the uniform linear array and other planar arrays, the UCA is moresuitable for applications such as radar, sonar, and wireless communications. On theotherhand, due to the180coverage in elevation plane, the spherical array is moresuitable for satellite commulication. In this dissertation, we mainly study thealgorithms of adaptive beamforming and DOA estimation for UCAs and sphericalarrays. Works and contributiones in the dissertation mainly include:(1) The mode transformation applied to the beamforming of coherent signals willcause the noise power altering. This leads to a degraded performance of theconventional beamforming algorithms. In order to solve this problem, we propose amethod of weighing virtual array of two concentric ring arrays, which significantlyimproves the output SINR.(2) The conventional beamforming algorithms of coherent signals usually havehigh computation complexity. An effective algorithm that uses real weights withoutphase shift is proposed. The interference is suppressed by using real weightgs and thedecorrelation of the signal sources is realized without spatial smoothing technique.Terefore the computation complexity is reduced. Moreover, this algorithm has rapidconvergency and stable performance in case of low SNR and finite snapshots.(3) A difference algorithm of DOA estimation for coherent sources with the UCA is proposed. Compared with the conventional DOA estimation algorithms, theproposed algorithm occupies stronger overload ability of signal sources and takes boththe resolution of the coherent and noncoherent signal sources into account.(4) A fast DOA estimation algorithm in the presence of an unknown mutualcoupling is proposed for UCAs. Meanwhile, a new method to remove the spuriousestimates is presented. Compared with the conventional DOA algorithms in thepresence of the unknown mutual coupling, the computation burden of the proposedalgorithm is greatly reduced.(5) A novel hybrid approach for2D DOA estimation in the presence of mutualcoupling based on sparse UCA is proposed. Different from the existing2D DOAalgorithms in the presence of mutual coupling, whose computation burdern is low, theproposed algorithm does not require the number of antenna elements to be sufficient toapply the traditional beamspace transoforamtion. Based on the manifold decompositiontechnique, a modified UCA-RARE algorithm is proposed to estimate the azimuth angleand the Root-MUSIC algorithm is employed to estimate the elevation angle. For sparseUCAs, compared with the original UCA-RARE, the modified UCA-RARE is able toavoid spurious estimates which only arise from the sparseness of the array elements.It's shown that the estimate accuracy usually depends on the truncation errorintroduced by manifold decomposition techniques. Hence, we analyze the truncationerrors for sparse UCAs and derive expressions describing the truncation errors in theDOA estimates.(6) Three different distributions of a spherical array antenna are introduced andtheir spherical mode transformations are discussed. A modified spherical ESPRITalgorithm is also presented. The performances of this algorithm applied to the threedifferent distributions are evaluated. Moerover, other two kinds of spherical ESPRITalgorithms are proposed. The performances of the proposed spherical ESPRITalgorithms are also studied.