An Improved Joint Sparse Representation of Array Covariance Matrices Approach in Multi-source Direct Position

The joint sparse representation of array covariance matrices (JSRACM) approach in direct position transforms the source position estimation problem into a spatial sparse signal representation (SSSR) optimization problem. With a novel binary sparse indicative vector (SIV) representing the support of joint SSSR of array covariance matrices, this SIVR-JSRACM algorithm presents high resolution, low computing complexity without knowing the number of sources and initial source positions estimates in advance. However, its performance degrades significantly along with the number of sources increases. To overcome this shortcoming, we proposed an improved joint sparse representation of array covariance matrices (IJSRACM) algorithm. The main contribution of this paper is that we estimate the elements in K-sparse covariance matrix of the potential sources. Thus we could get a new dictionary for SIV. The simulation results demonstrate that the SIVR-IJRSACM algorithm has superior localization accuracy and strong robust to noise under different numbers of sources and also remains the advantages of the SIVR-JSRACM algorithm listed above.