Efficient Wideband DOA Estimation Through Function Evaluation Techniques

This paper presents an efficient evaluation method for the functions involved in the computation of direction-of-arrival (DOA) estimators. The method is a combination of the Chebyshev and barycentric interpolators, and makes use of the discrete cosine transform. We present two applications of this method. The first is for reducing the complexity of the line searches in three wideband DOA estimators: incoherent multiple signal classification, test of orthogonality of projected subspaces, and deterministic maximum likelihood (DML). And the second application is a procedure to compress the wideband DML cost function, so that it is formed by just a few summands. This compression entails a reduction in complexity by a large factor. The evaluation method and its applications are numerically assessed in several numerical examples.