Robust Adaptive Beamforming with Ellipsoidal Steering Vector Uncertainty Set

The performance of adaptive beamforming degrades dramatically in the presence of steering vector uncertainties. In this paper, the robust adaptive beamforming with ellipsoidal steering vector uncertainty set is investigated. It belongs to the class of diagonal loading approaches. In order to select the loading level, the ellipsoidal uncertainty set is transformed into a spherical one first. Then the loading is determined based on worst-case performance optimization. Instead of being solved iteratively, the optimal loading is presented in closed-form here after some approximations. Compared with those iterative methods, the proposed one consumes less computational complexity and reveals how different factors affect the optimal loading. Numerical examples confirm the correctness and effectiveness of the proposed method.