Arthur B. Baggeroer
Depart. of Ocean Eng., Rm. 5-204, MIT, Cambridge, MA 02139
Matched-field tomography uses full wave modeling to infer geophysical parameters of the propagation medium. As such, it is fundamentally a problem in estimation theory. The signal processing literature has long used bounding methods to establish performance limitations and identify coupling among estimates of a set of parameters. The most extensive of these is the Cramer--Rao bound, which has been applied to the source localization problem for matched-field processing. The Cramer--Rao, as well as other bounds for analyzing the sidelobe, or ambiguity, problem can be applied to matched-field tomography. These will be derived and applied to problems in identifying the sound-speed profile and elastic properties of the seafloor.