Robust space-time scatter matrix estimator for broadband antenna arrays

In this paper we propose a robust method for estimation of the space-time array scatter matrix, i.e. the space-time pseudo-covariance matrix of the antenna array observations. Such scatter matrix and the subspaces spanned by its eigenvectors serve as a basis of broadband beamforming, high-resolution angle of arrival estimation methods and maximum likelihood array processing algorithms. The proposed estimator is based on a heavy-tailed probability model. Robustness and computational issues of the estimator are discussed. The considered estimator can be calculated using a simple iterative algorithm where no user-defined tuning constants are needed. The robust performance of the estimator is demonstrated in MVDR beamforming and GPS receiver examples, using a space-time uniform linear array (ULA). We demonstrate that the method is highly robust and suffers only a small performance loss in additive white Gaussian noise conditions.