Wideband Signal DOA Estimation By Using Sparse Representation Technique with An Acoustic Vector Sensor Array

Abstract

In order to improve the accuracy of wideband signal direction-of-arrival (DOA) estimation, an acoustic vector sensor array based DOA estimation method by using joint sparse representation and sparse reconstruction is proposed in this paper. According to the sparse signal decomposition theory, constructing the sparse representation model based on the wideband signal frequency-domain covariance matrix. And then exploiting the 12-norm to construct the joint sparse vector, which include the information of signal DOA, finally obtaining the estimated DOA results via the sparse reconstruction technique. The proposed method is insensitive to noise and doesn't need to estimate the number of signal sources. Simulation results display its superior DOA estimation performance.