Adaptive Beamforming in an Impulsive Noise Environment Using Matrix Completion

Abstract

In this letter, a new approach is presented for robust adaptive beamforming in an impulsive noise environment. In the proposed method, the Hampel identifier is first applied to identify and trim the entries contaminated by impulsive noise in the data matrix. Then, the trimmed data matrix is utilized to estimate the noise-free data matrix by exploiting its low-rank structure with the matrix completion technology. In order to cope with the errors introduced during the matrix completion process, the beamformer weight vector is obtained by optimizing the worst-case performance. Comparing with existing methods, the proposed approach does not require the priori information of the impulsive noise. Moreover, it recovers the contaminated data rather than simply discarding or normalizing them. In addition, errors caused by imperfect recovery are taken into account. By enjoying these advantages simultaneously, the proposed approach is capable of achieving a higher signal-to-interference-plus-noise ratio. Numerical examples are carried out to demonstrate its performance in different scenarios.