Angular-spectral decomposition beamforming for acoustic arrays

The method of principal component beamforming described in this paper is an array data reduction method that allows one to observe the statistically uncorrelated components of wave energy arriving at an array of acoustic sensors. The method can be used to process array data so as to observe and identify the sources of noise, both environmental and self noise. After identifying the sources of noise, the method of principal components can be used to discriminate signal from noise. The method can be applied to active systems (subbottom profilers) as well as passive systems. A model of isotropic noise and incident bandlimited plane waves is used to study array resolution and bandwidth effects. Experimental data from a2 \times 3planar acoustic array were used to identify sources of hydro-flow related noise in an underwater vehicle. In all cases studied, the technique provides a maximum spatial information analysis method to the observer.