Max-Min Beamforming for Multipath Exploitation

Abstract

This letter considers the robust spatial-filter design problem for multipath exploitation. The imprecise multipath return results in the output Signal-to-Interference-plus-Noise-Ratio (SINR) performance deterioration of traditional beamformers. We focus on enhancing the worst-case output SINR over the uncertain Direction-Of-Arrival (DOA) sector. To achieve robustness against covariance matrix errors caused by the desired signals contamination, we reconstruct the Interference-Noise Covariance Matrix (INCM) from Sample Covariance Matrix (SCM) based on the proposed residual noise components removal idea. Then, we cast the robust spatial-filter design problem as a max-min optimization problem, aiming at improving the worst-case output SINR. The filter vector is finally obtained via semi-definite relaxation followed by randomization synthesis. Simulation results demonstrate the effectiveness of the proposed algorithm.