Experimental Evaluation of a Null-Steered Performance Weighted Blended Beamformer

In array signal processing beamformers are used to improve the signal to interference and noise ratio (SINR). Dominant mode rejection (DMR) is a common adaptive beamforming algorithm designed to detect quiet sources in the presence of a few loud interferers. Buck and Singer proposed the Universal DMR (UDMR) beamformer that automates the selection of the subspace rank by blending fixed-rank DMR beamformers based on their performance. While UDMR requires only slightly more computation than a single fixed-rank DMR, it still requires a computationally expensive eigendecomposition. This paper proposes to reduce the computational complexity of the UDMR by replacing the DMR beamformers with a set of conventional beamformers. The null-steered performance weighted blended (NS-PWB) beamformer combines beamformers with nulls steered to interlocking locations so that if a plane-wave interferer is present at least one beamformer in the set will have a null nearby. Additionally, when there are persistent sources of non-plane-wave interference then beamformers can be designed and included in the beamformer set to reject those interferers when they are present. This paper applies the NS-PWB beamformer to data from a vertical line array in the Philippine Sea. The results show that the NS-PWB approximates the SINR performance of DMR beamformers with significant computational savings.