Probability-Constrained Robust Beamforming for General-Rank Signal Models

Adaptive beamformers such as LSMI provide an excellent robustness against various types of mismatches. These beamformers are sometimes too conservative because they are worst-case optimization based beamformers. A rank-one beamformer based on probability-constrained optimization has been proved to have better performance. In this paper, a probability-constrained beamformer for general-rank signal model is developed. The simulations have shown that this beamformer improves the Signal-to-Interference-plus-Noise Ration(SINR) of the system.