Broadband adaptive beamforming based on coherent signal subspace using spatial interpolation preprocessing

The coherent signal subspace (CSS) approach to broadband adaptive beamforming for correlated signal and interferences requires preprocessing by transformation matrices to focus signal subspaces at various frequencies into a common one. The transformation matrices often call for preliminary estimates of the directions of arrival of impinging signals through a spatial periodogram, and these preliminary angle estimates, if they are inappropriately selected, especially in the multigroup multiple signals, often result in an estimation error in the transformation matrices, and the array beamforming performance can be then degraded. Therefore an alternative method of preprocessing, called spatial interpolation, which does not require preliminary angle estimates, has been used instead of transformation matrices to estimate the arrival angles of signals impinging on a linear uniform array. The paper extends this idea to broadband adaptive beamforming. Performance comparisons of nulling coherent interferences using CSS with the two sorts of preprocessing are presented. Simulation results show the effectiveness of the spatial interpolation in adaptive beamforming