Steering vector estimation and beamforming under uncertainties

In this paper, we propose a new method for estimating the steering vector under uncertainties, which is utilized for improving the robustness of beamforming. We show that the desired steering vector can be estimated in closed form from a convex optimization problem by making use of the subspace principle. As this method is developed based on an extended version of the orthonormal PAST (OPAST), the steering vector can be recursively estimated with very low complexity and moving sources can be handled. To further improve the performance of beamforming, the uncertainty of the array covariance matrix is taken into account. Numerical results demonstrate that the proposed method performs well in the presence of uncertainties.