Phd thesis by Cristiano Soares, firstname.lastname@example.org
Abstract: Ocean Acoustic Tomography is a remote sensing technique that has been proposed to infer physical properties of the ocean traversed by the sound field. Although its feasibility has been demonstrated, it is still not being used in a systematic way due, in a large extent, to cost and operational difficulties of standard acoustic systems. Current developments of acoustic systems go in the sense of simplifying them, both at the emitting and receiving end. Simplifying an acoustic system may represent a loss or a reduction of the amount of information contained in the observed acoustic field, possibly conducting to degradation in the inversion results. The objective of this thesis is to adapt existing array processing methods to be used in acoustic tomography and geoacoustic inversion taking into account the challenges posed by such simplifications, and to cope with the loss of available information they may represent. Two aspects are exploited with the objective of coping with the reduction of information: one is the development of a broadband data model, and the other is the development of matched-field processors based on that broadband data model, with particular emphasis in high-resolution processors. Matched-field based approaches appear to be suitable to work in conjunction with the simplified acoustic systems used to collect several experimental data sets treated herein. Experimental results using simplified acoustic systems, sparse receiving arrays (active mode) on one hand, or an uncontrolled source (passive mode) on the other hand, show that it is possible to produce environmental estimates of the watercolumn and seafloor in close agreement with ground truth measurements.