Real-Valued Sparse Bayesian Learning For DOA Estimation

As one of the important research contents in the field of signal processing,direction of arrival(DOA)estimation has a wide range of applications in many fields,such as radar,sonar,mobile communication,etc.The traditional subspace methods(eg,MUSIC,ESPRIT)usually require a large number of snapshots to achieve excellent estimation performance,and when the signal is highly correlated or coherent,its estimation performance will be significantly reduced.In recent years,the emergence of sparse Bayesian learning(SBL)method has provided new ideas for DOA estimation.Scholars at home and abroad have also proposed many DOA estimation methods based on sparse Bayesian learning.However,the existing methods still have some shortcomings: 1)Most of the DOA estimation methods based on sparse Bayesian learning still require complex matrix operations in the process of solving signal parameters,which results in high computational complexity and large amount of calculation;2)Array flow pattern error is an important factor that affects the performance of DOA estimation.How to jointly estimate the signal parameters and array flow pattern error while reducing the computational complexity is also a problem that needs to be considered.In view of the above problems,this paper focuses on the real-valued sparse Bayesian learning for DOA estimation method,and the main research work of this paper is as follows:Aiming at the problem of high computational complexity and large amount of computation in the existing DOA estimation methods based on sparse Bayesian learning,this paper proposes an efficient real-valued spare Bayesian learning for off-grid DOA estimation method.The method uses the unitary matrix to transform the complex data model into a real one,and then uses singular value decomposition(SVD)to reduce the matrix dimension,which effectively reduces the computational complexity.However,after converting the data model to the real number domain,the most existing grid update methods will not be applicable.To solve this problem,this paper proposes a new grid update method.The new method uses a fixed stepsize to iteratively update the grid points,which can effectively improve the estimation accuracy.Experimental simulation results show that the method can effectively reduce the computational complexity and the amount of calculation,and at the same time achieve a higher DOA estimation accuracy.For the problem of how to effectively jointly estimate the array flow pattern error and signal parameters while reducing the computational complexity,this paper delves into the problem of joint estimation of DOA and array error calibration with mutual coupling between arrays,and proposes a joint estimation method of DOA and mutual coupling sparseness based on real-valued sparse Bayesian learning.In this paper,under the assumption of uniform linear array,Toeplitz mutual coupling coefficient matrix is introduced into the data model to establish a mutual coupling data model,and then the data model can be transformed into a real-valued model through unitary transformation.Finally the newly proposed real-valued spare Bayesian learning for off-grid DOA estimation method is used to estimate DOA.Experimental simulation results show that the method can effectively solve the problem of DOA estimation with mutual coupling between the arrays,and the estimation accuracy is good.