Two-dimensional Spatial Spectrum Estimation Algorithms Based On Reduced-dimension Technique

As one of the key issues in the array signal processing theory,the direction of arrival(DOA)of the target signal source can be obtained through spatial spectrum estimation the by measuring the signal energy distribution in space to locate the target signal source.Since the position of the target signal source in the three-dimensional space can be quantitatively characterized by its two-dimensional angle parameters,the two-dimensional spatial spectrum estimation is more practical in engineering projects.However,the commonly used methods based on twodimensional spatial spectrum search are often limited by the large amount of computation caused by the multi-dimensional spectrum peak search and rank deficiency of the covariance matrix caused by the spatial coherent signals.In this case,the thesis intends to solve the above problems from the weighted subspace fitting and compressive sensing theory and obtain twodimensional spatial spectrum estimation with lower complexity through dimension reduction process.The thesis includes the following parts:The work of this thesis is as follows:(1)Under the unified theoretical framework,two kinds of two-dimensional spatial spectrum estimation algorithms based on weighted subspace fitting are proposed respectively for uncorrelated signal sources and coherent signal sources.The key points are to parameterize the orthogonal projections of null space of two-dimensional unified array manifold,thereby transforming the optimization of two-dimensional weighted subspace fitting into two onedimensional versions.For uncorrelated signal sources,automatic pairing estimation of twodimensional angle parameters is achieved by optimizing one of the weighted subspace fitting function and the closed array manifold extraction technique.For coherent signal sources,the fitting functions of two one-dimensional weighted subspaces are respectively optimized to achieve the unpaired estimation of the two-dimensional angle parameters and then complete pairing through the uniqueness of the signal subspace.The proposed method has lower computational complexity and higher accuracy of angle estimation.At the same time,the algorithm for coherent signal sources does not need to be spatially smoothed,and the array aperture is effectively used.(2)The above method is performed on the basis that the type of signal source is known and has certain limitations.To solve the problem,a two-dimensional spatial spectrum estimation algorithm based on sparse signal reconstruction is proposed.The algorithm employs compressive sensing theory to decouple two-dimensional spatial spectrum estimation of uncorrelated/coherent signal sources into the reconstruction of two one-dimensional sparse signal,which greatly reduces the computational complexity of two-dimensional sparse signal reconstruction.Then pairing angle through the uniqueness of the signal subspace.Since the sparse reconstruction problems were reduced to two one-dimensional reconstruction,the proposed algorithm has lower computation complexity.