Research On High Resolution Direction Finding Of Spatial Spectrum Based On Array Reconstruction

Spatial spectrum estimation is the core research area of array signal processing theory,especially with the gradual and rapid evolution of mobile wireless communication technology,it is widely used in areas such as intelligent connected cars,drones,and unmanned boats to achieve multi-intelligence Functions such as location awareness and cooperative transmission.This thesis focuses on the direction-of-arrival(DOA)parameter estimation in spatial spectrum estimation theory,and focuses on the specific implementation of matrix reconstruction technology in two different signal model scenarios.For the scene where uncorrelated and coherent signals are mixed,the weighted subspace fitting technique and oblique projection technique are used to reconstruct the array observation matrix to separate the two signals,and then an efficient two-dimensional DOA algorithm can be obtained.In the source scene,the spatial redundancy of the two-dimensional area array is used to reconstruct the array observation matrix,and the center angle and expansion angle of multiple signal sources are estimated.The main work of this paper is as follows:(1)Under the point source model,for the mixed signals with uncorrelated and coherent signals,the weighted subspace fitting is used to implement the dimensionality reduction processing of the two-dimensional estimation,which is transformed into a one-dimensional angle estimation problem.To obtain the DOA information of mixed signals by extracting the corresponding information from the array manifold to realize the automatic pairwise estimation of the two-dimensional angle of the uncorrelated signal source;then a new oblique projection algorithm is designed to convert the information is separated and a new array observation matrix is reconstructed to obtain the two-dimensional angle information of the coherent signals.Finally,the angle information obtained before the local search optimization is used to finally obtain an efficient algorithm.(2)In the case of multiple distributed source signals coherent with each other,the traditional subspace decomposition method will destroy its original integral structure after smoothing,and applying a smoothing algorithm to the model cannot decoherence,resulting in an inability to estimate the extended angle.Therefore,in this paper,a distributed model of the integral structure is replaced by an approximate model of Schur-Hadamard product,and the spatial redundancy of the two-dimensional area array is used to reconstruct the array observation matrix,thereby solving the problem that the extended angle cannot be estimated.The central angle of arrival and the extended angle can be solved simultaneously.