Research On High Degree Of Freedom DOA Estimation Technology Based On The Coprime Array

Multi-target DOA estimation technology is widely used in military and civilian fields.The multi-target positioning method based on traditional uniform array is relatively mature.In recent years,the copime array structure has attracted extensive attention from experts and scholars because of its advantages of larger equivalent aperture,higher degree of freedom,weaker mutual coupling effect and higher estimation accuracy.At present,the DOA estimation methods based on the coprime array are mostly conducted by ulilizing the covariance information.For Gaussian signals,the first-order and second-order statistices can fully describe their statistical characteristics,and the value of high-order statistic is zero.However,most of current communication signals are non-Gaussian,and their high-order statistics contain more valuable information,which can be used to further improve the estimation performance.In view of the above aspects,this thesis studies the DOA estimation algorithm based on the coprime array model,with the non-Gaussian signal target,and the estimation freedom of the algorithm is taken as the research goal,from the perspective of the fourth-order cumulantbased array expansion,non-circular information and the array error.The main contents are as follows:1.High degree of freedom DOA estimation algorithm under the coprime array for nonGaussian signals.The algorithm takes advantage of the fact that the fourth-order cumulant of nonGaussian signals is not zero.Firstly,the fourth-order cumulant matrix of the array received signal is constructed,and then its constituent components are analyzed and mapped into a virtual array form.Next,to reduce the computational complexity,the continuous response part is selected to construct a continuous virtual uniform array.Finally,the fast subspace class algorithm can be applied to solve the DOA estimation value.Compared with the classical high-degree DOA estimation method based on vectorization reconstruction,the proposed algorithm has twice the degree of freedom.Therefore,in the non-Gaussian signal target scenario,the proposed algorithm could effectively increase the number of estimable targets.2.High degree of freedom DOA estimation algorithm for non-Gaussian signals with non-circular features under the coprime array.The algorithm constructs an array receiving model with non-circular features signals under the coprime array,and a fourth-order-cumulantbased high-degree-of-freedom DOA estimation algorithm combined with non-circular features under a coprime matrix is proposed.The main idea of the method is to further expand the array aperture by using the non-circular information possessed by the non-circular signal.Firstly,according to the different definitions of the fourth-order cumulant matrix,an extended fourth-order cumulant matrix is constructed.According to its equivalent virtual array form,the expression of its spectral peak search is given.Then,to reduce the computational complexity,the virtual array is reconstruct and the ESPRIT is used to directly solve the DOA estimation.The simulation results show that the estimated number of targets of the proposed algorithm can reach the degree of theoretical analysis,and the estimated degrees of freedom and estimated resolution are further improved compared with the previous estimation methods for general non-Gaussian signals.3.High-degree-of-freedom DOA estimation algorithm for real non-Gaussian signals under under the coprime array.The real signal is a kind of non-circular signal,which is a completely non-circular signal with zero initial phase.Combined with this feature,this thesis proposes a high-degree-of-freedom DOA estimation algorithm for real non-Gaussian signals under the coprime array.The algorithm utilizes the characteristics that the virtual sensor and the original array sensor have the consistent response when the initial phase of the real non-circular signal is zero.The vectorization reconstruction method is used to further expand the virtual array aperture,combined with the spatial smoothing technique and the subspace method.The simulation results verify the consistency of the estimable target number and the theoretical analysis degree of freedom of the proposed algorithm.The estimated freedom and resolution of the real non-circular signal are further improved.4.Non-Gaussian DOA estimation method with sensor's amplitude and phase error under coprime Array.The method realizes error correction by compensating on the received data,and realizes a rough estimation of the DOA parameter of the non-Gaussian signal;On this basis,in order to further improve the estimation accuracy and consider the existence of a certain position error of the actual correction source,a self-correcting iterative fine correction algorithm is proposed.The algorithm obtains a more accurate DOA parameter estimation value by setting the rough estimation result of the DOA parameter to the initial value and correcting it by multiple iterations.Simulation experiments show that the proposed error correction algorithm can accurately calibrate the error and achieve a high-precision DOA estimation.