An improved SVD based algorithm for bearing estimate

A study of the direction of arrival (DOA) matrix is presented. For the case of low signal-to-noise ratio (SNR) the standard DOA matrix usually yields poor performance. By means of the singular value decomposition (SVD) technique, with a cleaned or SNR enhanced data record, the author propose construction of the DOA matrix by the lower-rank Froebinius approximation. Two types of improved DOA matrices are given, and two improved algorithms for 2D angles of arrival are derived. Simulation results are provided.<
>