Sector-focused processing for stabilized resolution of multiple acoustic sources

Abstract

Sector-focusing (SF) is a technique for performing stable high resolution acoustic matched-field processing (MFP) in a waveguide; it has been shown to be highly effective against sensor phase error and acoustic mismatch in an environment dominated by modal noise [Frichter et al., J. Acoust. Soc. Am. 88, 2843–2851 (1990); Smith et al., J. Acoust. Soc. Am. 93, 2617–2626 (1993)]. SF is successfully applied here to a problem of critical importance in MFP: The resolution of multiple, closely spaced acoustic sources under severe model parameter mismatch conditions. The mechanism for mismatch induced instability in Capon’s estimator [Proc. IEEE 57, 1408–1418 (1969)] is described, and the SF procedure for eliminating the resultant sidelobe ambiguities is developed. By judicious sector design and dimensionality selection, these sidelobes are reduced to unambiguously resolve the sources present in a waveguide. Simulations were performed for multiple cw sources in shallow water. It was found that SF can always be made to achieve the stability and performance of the Bartlett processor in the ‘‘small’’ sector limit. In the ‘‘large’’ sector limit, SF is equivalent to the reduced maximum likelihood (RML) processor [Byrne et al., J. Acoust. Soc. Am. 87, 2493–2502 (1990)]. By careful sector selection, the method can be readily tuned to achieve better performance in the intermediate regime.