Source localization in noisy and uncertain ocean environments

Abstract

Interference from noise and uncertainties in the environmental parameters are arguably the two most serious limitations in matched-field processing (MFP). Among the techniques that have been developed for handling these difficulties are the noise-canceling processor [M. D. Collins, N. C. Makris, and L. T. Fialkowski, “Noise cancellation and source localization,” J. Acoust. Soc. Am. 96, 1773–1776 (1994)] and focalization [M. D. Collins and W. A. Kuperman, “Focalization: Environmental focusing and source localization,” J. Acoust. Soc. Am. 90, 1410–1422 (1991)]. The noise-canceling processor is a generalization of the Bartlett processor that is based on matching the covariance matrix of the data with replica covariance matrices of the signal and the noise. Simulations are presented to illustrate the performance of the noise-canceling processor when there are errors in the noise replica. Focalization is a generalization of MFP in which environmental parameters are included along with source parameters in the search space. An implementation of this approach that is suitable for applications is developed and tested. The noise-canceling processor and focalization are used to simulate the localization of a source buried in noise in an uncertain environment.

Source localization in the presence of gross sediment uncertainties

“…Low signal-to-noise ratio can be handled by including an estimate of the covariance of the noise in the processing. 10 In this paper, we are concerned with the issue of uncertain sediment parameters. Since small amounts of uncertainty can cause matched-field processing to fail and there are usually gross uncertainties in the sediment, it was not obvious that focalization would work for this problem before we actually tried it.…”

Experimental testing of the noise-canceling processor

“…7 The noise-canceling processor has also been tested in simulations for problems involving environmental uncertainties. 8 The noise-canceling processor is based on matching the data covariance matrix with replica covariance matrices for the signal and the noise. The conventional approach for obtaining replicas is to solve the wave equation; an alternative approach, which circumvents the need for information on the environmental parameters, is to use acoustic field data.…”

General second-order covariance of Gaussian maximum likelihood estimates applied to passive source localization in fluctuating waveguides

“…Incomplete or imprecise knowledge of environmental parameters and randomness in the propagation environment are known to seriously deteriorate the performance of MFP, which has been investigated extensively in the past. [9][10][11][12][13][14][15][16][17][18] MFP has been demonstrated in a number of theoretical and experimental scenarios involving fluctuating or unknown environments, 14,17,18 but with significant localization ambiguities due to multimodal propagation and environmental mismatch. For example, in Ref.…”