DOA estimation method for co-arrays with unknown number of sources

Abstract

In the present paper, we consider a co-array as a coprime array or a nested array. Pal et al. proposed a method to extend a co-array to a larger virtual array, then implemented the spatial smoothing technique to construct the covariance matrix of a virtual uniform linear array (ULA). Thus subspace-based direction of arrival (DOA) estimation algorithms can be used to detect more sources than the number of array elements. However, since the subspace-based DOA estimation methods are applied, the DOA estimation accuracy depends on the performance of source number estimation. By employing a set of Toeplitz matrices, we propose a DOA estimation method for co-array, which does not need to know the number of sources prior to computing the spatial spectrum. Computer simulations are provided to demonstrate effectiveness of the proposed approach.