A Distributionally Robust Minimum Variance Beamformer Design

Abstract

IEEE This paper is concerned with a robust minimum variance beamformer design. To hedge the mismatch between the true and the assumed steering vectors, a distributionally robust beamformer (DR-beamformer) is proposed. The tractable reformulation of this beamformer is developed. Compared with the existing robust beamformers (worst-case robust beamformer, Gaussian robust beamformer), the proposed robust beamformer does not assume full knowledge of the channel mismatch. Therefore, it is more flexible in practice and more general in formulation. In addition, the relationships of the proposed robust beamformer with the existing ones are investigated. The performance gain of the DR-beamformer over the other robust beamformers is highlighted through numerical simulations.