Robust Adaptive Beamforming Based on Subspace Decomposition, Steering Vector Estimation and Correction

Abstract

Considering that the performance of adaptive arrays is sensitive to any type of mismatches, an innovative robust adaptive beamforming method based on covariance matrix reconstruction, subspace decomposition, steering vector estimation and correction is proposed. Based on Capon spatial spectrum, a group of angle sets containing all interfering signals are determined, and the interference covariance matrix can be reconstructed with a smaller integration interval. On the other hand, the sample covariance matrix can be decomposed into signal subspace and interference-plus-noise by using the principle of maximum correlation. Based on the interference-plus-noise subspace and the reconstructed signal-plus-noise covariance matrix, a new convex optimization model is built to estimate the steering vector of the desired signal. Then, an improved projection approach based on signal subspace is designed for correction to improve the robustness against the nominal direction vector mismatches. Simulation results demonstrate that the proposed method achieves better overall performance under multiple mismatches over a wide range of input signal-to-noise ratios.