Robust adaptive beamforming based on multi-dimensional covariance fitting

Robust adaptive beamforming based on worst-case performance optimization is known to provide a substantially improved robustness against signal self-nulling as compared to the traditional adaptive beamforming techniques. The worst-case performance optimization based beamformers of and can be alternatively obtained by solving the one-dimensional (1D) covariance fitting problem. In this paper, we show that the robustness of this approach can be significantly improved by extending it to multi-dimensional (MD) covariance fitting.