Research On Direction Of Arrival Estimation Based On The Coexistence Of Uncorrelated And Coherent Signals

Direction of arrival(Direction of Arrival,DOA)estimation of incident signals is one of the research hotspots in the field of array signal processing.Spatial spectrum estimation algorithms are widely used in radar,sonar,detection,medical and many other fields for their excellent DOA estimation performance.However,the premise of the good angle estimation ability of this kind of algorithm is that the incident signal is uncorrelated.However,in the actual environment,the mixed signal formed by the mixture of uncorrelated signal and coherent signal is common.In this case,the rank of the covariance matrix received by the array will be deficient,and the angle estimation ability of the spatial spectrum estimation algorithms of these subspace decomposition classes will be greatly reduced.Therefore,the research on DOA estimation of mixed signals not only has theoretical significance,but also has certain engineering realization value.This paper makes an in-depth study on the DOA estimation of mixed signals.This paper puts forward the corresponding solutions and improvement schemes for some problems existing in the existing algorithms,including the following aspects:1.Firstly,the problem of two-dimensional DOA estimation with the coexistence of uncorrelated and coherent signals under scalar array is studied,and a mixed signal two-dimensional DOA estimation algorithm based on multiple matrix reconstruction is proposed.Considering the mixed signal jointly,four virtual covariance matrices are constructed by using the autocorrelation vector and cross-correlation vector of the data received by the double parallel line array.the four virtual covariance matrices and their backward versions are further reconstructed to get a joint matrix.Using the idea of propagator algorithm,the propagator corresponding to the joint matrix is calculated.The propagation operator is extended to obtain the orthogonal subspace which is orthogonal to the virtual steering vector,and then a pseudo-spectral function is constructed,and the one-dimensional DOA estimation result of the mixed signal is directly obtained by finding the root of the pseudo-spectral function.After obtaining the one-dimensional DOA information,a new virtual covariance matrix is constructed by using the cross-correlation vectors between the double parallel linear arrays,and a new joint matrix is obtained by combining the covariance matrix with the original virtual covariance matrix.The propagator of the new joint matrix is calculated again,and the propagator is extended to a new form.Combined with the previously obtained one-dimensional DOA information,the estimation of another dimensional angle is realized and automatic pairing is realized.The proposed algorithm can realize the two-dimensional DOA estimation of mixed signals with very low computational complexity.Computer simulations verify the effectiveness of the proposed algorithm.2.Compared with the traditional scalar sensor,the joint estimation of DOA information and polarization parameters of uncorrelated and coherent signals under polarization sensitive array is studied.Based on the orthogonal dipole polarization sensitive line array,a joint estimation algorithm of one-dimensional DOA information and polarization parameters of uncorrelated and coherent signals is proposed.Based on dual parallel COLD array,a joint estimation algorithm of two-dimensional DOA information and polarization parameters for uncorrelated and coherent signals is proposed.For the joint estimation algorithm of one-dimensional DOA information and polarization parameters,firstly,the target matrix is constructed,the incident angle of the unrelated signal is estimated according to the subspace method,and the polarization parameter of the unrelated signal is estimated by using the correlation of the data received by the orthogonal dipole array.In the process of estimation,the angle information and polarization information can be matched correctly.Then the eigenvectors are rearranged to estimate the incident angle of each coherent group,and the polarization parameters of each coherent group are estimated by the least square method.Compared with the existing algorithms for all-electromagnetic vector sensors,the proposed algorithm reduces the consumption of hardware resources,realizes the estimation of polarization information,and increases the number of estimable mixed signals.Compared with the existing similar algorithms,the proposed algorithm improves the angle estimation accuracy of coherent signals.Simulation results verify the effectiveness and reliability of the proposed algorithm.For the joint estimation algorithm of two-dimensional DOA information and polarization parameters,firstly,the covariance matrix of the received data of the array is calculated,and the signal subspace corresponding to the guidance vector of the incident signal is obtained by eigendecomposition of the covariance matrix.The idea of ESRPIT algorithm is used to estimate the one-dimensional DOA information of unrelated signals,and then the polarization parameters of unrelated signals are estimated by using the correlation between the received data of COLD array.Combined with the estimated one-dimensional DOA information and the signal subspace of the whole double parallel array,the other dimensional angle is estimated and the correct pairing is realized at the same time.After the estimation of uncorrelated signals is realized,the feature vectors corresponding to each coherent group are rearranged to construct three virtual covariance matrices corresponding to each sub-array.The three virtual covariance matrices contain the two-dimensional DOA information and polarization parameters corresponding to each coherent group,and ensure that the rank of the three virtual covariance matrices is restored in the process of construction.According to the idea of ESPRIT algorithm,these three virtual covariance matrices can be used to estimate the one-dimensional angle information of the signals contained in the current coherent group.By further decomposing the three virtual covariance matrices,it can be found that the polarization parameters and another one-dimensional DOA information can be solved by the idea of least square,and then the DOA information and polarization parameters of the mixed signal are obtained.