Study On Spatial Spectrum Estimate Algorithm With Sparse Array

Direction of arrival(DOA)estimation of multiple incident signals is one of the main issues in array signal processing field.Many traditional DOA estimation algorithms are based on the data model of uniform array.However,in order to avoid angle ambiguity,the inter-senor spacing of uniform array must be less than half wavelength of the incident signals.So,the aperture of uniform array is limited to the number of sensors.Unlike uniform array,the inter-senor spacing of sparse array can be different and break the limit of half wavelength of the incident signals.Therefore,under the same number of elements,compared with uniform array,sparse array can perform better in expanding array aperture,improving accuracy and enhancing resolution.Just because of this,in recent years,sparse arrays have received more and more attention by scholars.In this dissertation,our research focuses on the application of sparse array in DOA estimation,and the main contents are summarized as follows:Firstly,we study the application of sparse array in one-dimensional(1D)DOA estimation.Above all,a low-complexity multiple signal classification(MUSIC)algorithm based on a coprime array is proposed.The advantage of this algorithm is that the fourth-order-cumulants(FOCs)of received data are used directly to construct an extended signal subspace.So,this algorithm not only doesn’t need to construct a high-order covariance matrix to expand the aperture of array,but also doesn’t need to implement eigenvalue decomposition(EVD)to obtain signal subspace.Furthermore,we can reduce the search scope of the MUSIC algorithm by a pre-estimation technology.Compared with the traditional MUSIC and FOC-based MUSIC(FOC-MUSIC)algorithms,this algorithm shows higher resolution and accuracy.Then,a blind DOA estimation algorithm with a two-level nested array under unknown mutual coupling is presented.In this algorithm,a two-level nested array is used to reduce the effect of mutual coupling.Using the FOCs of received data from partial sensors to construct a FOC matrix that is equivalent to a covariance matrix of the received data from a uniform array.By dealing with the FOC matrix,we can get higher-resolution and higher-accuracy DOA estimation results.Simulation experiments show that the performance of this algorithm is better than many existing DOA estimation algorithms under unknown mutual coupling.Secondly,we study the application of parallel sparse arrays in two-dimensional(2D)DOA estimation.Above all,two array manifold matching approaches are proposed with parallel linear arrays for the DOA estimation of incoherent and coherent signals,respectively.The azimuth angles can be fast estimated after the elevation angles are estimated by some existing one-dimensional(1D)DOA estimation algorithms.Particularly,the azimuth angles are automatically paired with the elevation angles.Then,two 2D DOA estimation algorithms based on two parallel two-level nested arrays are proposed.For the two methods,FOCs of received data are used to construct two FOC matrices,by dealing with one of which we can get the estimation of elevation angles.By using the estimated elevation angles,two methods are driven to estimate the azimuth angles without any peak search and eigenvalue decomposition(EVD).Moreover,the two algorithms can achieve the automatically paired 2D DOA estimation.Thirdly,we study the application of L-shaped sparse arrays in 2D DOA estimation.Above all,a 2D DOA estimation algorithm with an L-shaped coprime array is proposed.The angle estimation problem is addressed by arranging the FOCs of array received data to form a FOC matrix which is equivalent to a cross-covariance matrix of the received data from a uniform L-shaped array.The estimation accuracy of this algorithm is better than many existing classical algorithms based on uniform L-shaped array.Then,a subspace expansion technique with an L-shaped two-level nested array and a matching algorithm for arbitrary L-shaped array are presented.The biggest advantage of this subspace extension technique is that it can obtain an extended signal subspace with small amount of computation.Using the subspace extension technique,we can improve the estimation accuracy of elevation and azimuth angles.Combining the new pairing method,we can obtain the last 2D DOA estimations.Fourthly,we study the application of sparse array in DOA estimation of non-circular signals.Firstly,two 1D DOA estimation algorithms for non-circular signals with a coprime array is proposed.For the two algorithms,by using the received data from the array,a uniform method is introduced to construct four FOC matrices,which can be combined into a high-order partitioned matrix.Then,using the principles of MUSIC algorithm for non-circular signals(NC-MUSIC)and ESPRIT for non-circular signals(NC-ESPRIT),we can get high-performance DOA estimation results.Simulation results show that the estimation performance of the two algorithms is better than some existing algorithms.Besides,a 2D DOA estimation algorithm for non-circular signals with an L-shaped coprime array is proposed.Like the 1D DOA estimation algorithm of non-circular signals,we also need to construct four FOC matrices and a high-order partitioned matrix.By dealing with the high-order partitioned matrix,we can obtain the estimations of automatic pairing elevation and azimuth angles.Simulation results show that the estimation performance of this algorithm is better than some existing algorithms.At last,we study the application of array manifold matching algorithm and subspace extension algorithm in bistatic Multiple-Input Multiple-Output(MIMO)radar.Above all,a fast direction finding algorithm in bistatic MIMO radar is proposed with array manifold matching technique.This algorithm can reduce the complexity and obtain automatically paired direction of departure(DOD)and direction of arrival(DOA).Then,the subspace extension technique based on two-level nested is applied in bistatic MIMO radar and a new joint estimation algorithm for DOD and DOA is proposed.This algorithm can improve the estimation accuracy of subspace-based algorithms with small amount of extra computation.