Abstract:
This paper describes array processing procedures which have been developed to investigate the ambient noise in the ocean. Most previous studies of ambient noise have been concerned with its detrimental effect on array performance. More recently however the ambient noise is being used to study oceanographic features and locate passive objects. Array processing methods which have been developed for active imaging or passive narrow-band detection are not necessarily optimal for this application. In this paper procedures for imaging using very broadband signals are described and it is shown that good resolution can be obtained using a volume array with a small number of transducers (four to six), by processing the cross-correlation matrix of the signals. Resolution is enhanced by an order of magnitude by using a deconvolution procedure combined with a Hilbert transform. The optimum processing procedure is found to be a multiplicative method which only uses the cross correlations between a reference transducer and the others in the array. Experimental results are shown which verify the predicted array performance. [Work supported by ONR.]