Covariance Denoising with Applications to Adaptive Beamforming

Adaptive beamforming is of high importance in array signal processing applications that require interference suppression. In practice, minimum variance distortion-less re-sponse (MVDR) is one of the most widely used algorithms for adaptive beamforming. Classical MVDR implementations require processing number of samples larger than the number of elements in the array, whereas existing sample-based methods have high computational complexity. This work introduces computationally efficient sample-based method for MVDR beamforming. Utilizing knowledge of the array manifold, the method first finds a sparse representation of the covariance, which is then used to compute MVDR beamformers. The proposed method has low sample complexity and performs well as long as the number of ele-ments is much greater than number of interfering sources. This feature makes the method a promising candidate for adaptive beamforming with large arrays.