Matched field processing: Source localization in correlated noise as an optimum parameter estimation problem

Matched field processing is a parameter estimation technique for localizing the range, depth, and bearing of a point source from the signal field propagating in an acoustic waveguide. The signal is observed at an array in the presence of additive, spatially correlated noise that also propagates in the same ocean environment as the signal. In a weak signal-to-noise situation this parameter estimation requires the maximum exploitation of the physics of both the signal and noise structure which then must be coupled to optimum methods for the signal processing. We study the physics of this processing by modeling the ocean environment as a waveguide that is horizontally stratified with an arbitrary sound-speed profile in the vertical. Thus, the wave equation describes the underlying structure of the signal and noise, and the signal processing via the generation of the replica fields. Two methods of array processing are examined: (i) the linear cross correlator (Bartlett) and (ii) the maximum likelihood method (MLM) for the parameter estimation procedure. The optimum potential resolution is evaluated using a generalized Cramer–Rao bound. The two processing methods and the lower bound demonstrate that the ability to reject ambiguities is determined not only by the signal-to-noise ratio but also by the relative spatial structures of the signal and noise. Simulations of both the array processing methods and the bounds for shallow water and Arctic environments using full wave modeling of the signal and noise fields illustrate the coupling of the ocean environment to the localization performance.