High resolution direction of arrival estimation of multiple wide-band sources in multichannel adaptive nulling systems

The main purpose of this paper is to demonstrate that the adaptive weights obtained from different frequency channels can be used for high-resolution direction of arrival (DOA) estimation. It is shown that these weights have sufficient information which can be processed by eigenstructure methods to yield the signal and the noise subspaces. As such, with the availability of the adaptive weights in a multichannel nulling system, source localization using the data matrix or time-averaged estimation of the covariance matrix may, in most casts, prove unnecessary and can only add meager improvement over the proposed method. We present two approaches to extract the eigenstructures from the multichannel adaptive weights. Both approaches are based on coherent subspace averaging, namely the focusing techniques