Robust adaptive beamforming method for large-scale array with automatic diagonal loading and steering vector estimation

Here, the authors propose a novel robust adaptive beamforming method, which is robust to both of the covariance matrix mismatch and the steering vector estimation mismatch. Their method was based on a recently proposed adaptive beamforming technique for large-scale arrays. In this beamforming technique, a sparse sub-array having much smaller number of antenna elements is first selected from the original physical array to perform digital sampling, and then the beamforming coefficients for the original full array are calculated by using only the digital samples from the selected sub-array. In their robust method, they calculate the diagonal loading level automatically by using the spatial-matched method. Then, the steering vector of the desired signal is estimated by maximising the beamformer output power under a constraint that the estimated steering vector does not converge to any of the interference steering vectorsor their linear combinations. Numerical examples are provided to demonstrate the effectiveness of the proposed method against covariance matrix mismatch and steering vector error.