Abstract
Matched-field processing (MFP) for source localization can be highly sensitive to mismatch between the true bathymetry and the bathymetry model used to compute the replica fields. This paper presents an efficient algorithm to invert for source location and an optimal representation of the (range-dependent) bathymetry which is applicable whenever the adiabatic normal-mode model applies. In the adiabatic mode formulation, the range dependence of the bathymetry between source and receiver is completely contained in range integrals of the wave numbers. Parametrizing these integrals by a single-pole expansion maps all bathymetries which have an equivalent effect on the propagation into as little as a single parameter, resulting in a tremendous reduction in the dimensionality of the inverse problem. The resulting parameter space can then be searched for the bathymetry parameters which provide the best match with the measured acoustic fields. This search can be carried out very efficiently, and applied to optimize the bathymetry for each grid point of an ambiguity surface. The optimization procedure has the effect of broadening the peaks of the ambiguity surface. This results in a loss in resolution, but allows the use of a coarser search grid.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call