Robust algorithm against large look direction error for interference‐plus‐noise covariance matrix reconstruction

Abstract

A novel robust algorithm for the interference-plus-noise covariance matrix (IPNCM) reconstruction according to its definition is addressed. Unlike the previous works, each interference steering vector (SV) is estimated as the prime eigenvector of the interference covariance matrix obtained by integrating over each interference angular sector and its power through the Capon spectral estimator is also estimated. Then, a more precise IPNCM based on these accurate estimates is reconstructed. In addition, the same idea is used to estimate the SV of the signal of interest over the signal angular sector. Simulation results demonstrate that the proposed beamforming has much more robustness against the large look direction errors of the interferences, and can attain the optimal beamformer in both high and low signal-to-noise ratio cases with low computational complexity.