Generalized principal component beamformer for communication systems

The contribution of this paper is two-fold. First, we introduce a generalized principal component (GPC) beamforming technique for reduced rank processing that allows a trade-off between interference and noise reduction via the introduction of a control parameter, ε. Three variants of the GPC beamformer corresponding to ε=1 (called T1 beamformer), ε=0.5 (called T2 beamformer) and ε=0 (called T3 beamformer), which maximize the signal-to-interference ratio, the signal-to-interference plus noise ratio, and the signal-to-noise ratio at the array output respectively, are considered in detail. The second contribution of this paper is to compare the robustness between the reduced rank and full rank beamformers. We use analytical studies and computer simulations to show that the T2 and T3 beamformers are robust against calibration and/or pointing errors.