Eigenvector adaptive noise suppression applied to vector sensor arrays.

Due to the characteristics of acoustic intensity beamforming, it is desirable to reduce the noise level of the measured spectrum. In this paper, eigenvector noise suppression is adaptively applied to a vertical line array of underwater acoustic vector sensors. Eigen decomposition is applied to the cross‐spectral density matrix (CSDM) of the beamformed complex spectrum from an array of Wilcoxon VS‐206 vector sensors. The array elements are integrated sensors with a triaxial accelerometer and an omni hydrophone. The beamformer steering vector is applied to each of the eigenvectors for a range of steering directions within the noise field of interest. Using a known target bearing, the signal‐to‐noise ratio (SNR) is computed for the beamformed eigenvectors as the frequency averaged ratio between the target steering direction and all other steering directions. The eigenvectors with an SNR below a certain level are determined to be dominated by interference. These eigenvectors are subtracted from the CSDM using a projection matrix to suppress the interference. The results using calibrated source level data are presented. [Work supported through ILIR grant from Carderock Division, NSWC.]