On the Robustness of Covariance Matrix Shrinkage-Based Robust Adaptive Beamforming

Beamforming has been widely studied in wireless communications, radar, sonar, and other array systems. Digital beamforming is usually designed based on the array response and the estimation of the covariance matrix of the received signal. Various model mismatch issues can arise due to unequal antenna gains, phase errors, direction-or-arrival (DOA) mismatch and imperfect estimation of the covariance matrix. Different methods based on the shrinkage estimation of the covariance matrix and interference-plus-noise covariance matrix reconstruction have been proposed to address the challenges. In this paper, we investigate the robustness of several approaches in the presence of model uncertainties. We demonstrate the pros and cons of those approaches under different scenarios, based on which recommendation on the choice of the proper method may be made.