Research On Two-Dimensional DOA Estimation Methods Based On Sparse Array

Direction of arrival(DOA)estimation is the main research content of array signal processing,which has a wide range of applications in communication,radar,sonar and other fields.Sparse array has been widely studied in DOA estimation due to the advantages of array aperture.In this paper,the classical two-dimensional DOA estimation methods are analyzed,and the sparse array is applied to L-shaped array and circular array to solve the problems of estimation accuracy and estimation cost.Then four kinds of sparse array structures and corresponding two-dimensional DOA estimation methods are studied,which effectively solve the related problems.The main research contents are as follows:(1)The two-dimensional DOA estimation methods based on L-shaped array are introduced,including Multiple Signal Classification(MUSIC)method,Estimation of Signal Parameter via Rotational Invariance Techniques(ESPRIT)method and Propagator Method(PM)method.The three methods have the problems of estimation accuracy and computational complexity in twodimensional DOA estimation,so it is necessary to study new array structures and corresponding two-dimensional DOA estimation methods to solve the related problems.(2)A two-dimensional DOA estimation method based on nested L-shaped array is studied.Firstly,two signal subspaces are obtained by singular value decomposition of cross-covariance matrix of received date.Then,one extended signal subspace is constructed by two rotation factors.Finally,the ESPRIT method is used for angle estimation and the corresponding pairing method is used for angle pairing.The method does not obtain the extended signal subspace by constructing extended covariance matrix,and it has higher estimation accuracy and lower computational complexity.(3)A two-dimensional DOA estimation method based on coprime L-shaped array is studied.Firstly,one fourth-order cumulant matrix is constructed by fourth-order cumulant of received data.Then,two DOA matrices are constructed.Finally,the DOA matrix method is used to obtain angle estimations by eigenvalue decomposition of the two DOA matrices.The method uses the fourth-order cumulant matrix instead of covariance matrix to construct the DOA matrices,which can suppress gaussian noise and realize automatic angle pairing,and it has higher estimation accuracy.(4)A two-dimensional DOA estimation method based on nested circular array and coprime circular array is studied.Firstly,one received data of new virtual array is obtained by covariance matrix of received data.Then,the compressed sensing method is used for angle estimation.Finally,one covariance matrix of virtual uniform circular array is constructed by using the matrix filling method,and the compressed sensing method is used for angle estimation again.The method achieves similar high estimation accuracy for different angles by structural characteristic of uniform circular array.A coprime circular array with coprime element spacing is designed,which can reduce the angle dependence and improve the estimation accuracy by more uniform element spacing.