Title:  Listening to the ocean: solving inverse problems in underwater acoustics  

Speaker: Prof. Zoi-Heleni Michalopoulou (Professor and Chair of Department of Mathematical Sciences, New Jersey Institute of Technology)

Abstract:  Source localization and geoacoustic inversion are of utmost importance in antisubmarine warfare and environmental monitoring.  Matched-field processing is a method often employed to address these tasks.  Matched-field inversion (MFI) is the process of maximizing a correlation measure between received data at an array of hydrophones and replica fields calculated with a sound propagation model for a set of values for environmental parameters as well as source range and depth. The values that produce the maximum correlation form the geoacoustic parameter and source location estimates.  In this presentation we discuss the mechanics of MFI and we demonstrate improvements in performance by using Gaussian Process regression; this enables dense sampling of the water column by introducing the concept of virtual receivers.  Although MFI can be shown to be, in general, highly successful, it is, under certain circumstances, computationally intensive and other approaches become more desirable for addressing inverse problems in ocean acoustics.  We will show cases when it is preferable to view the problem from a different perspective and we will discuss an alternative methodology, relying on linearization and sequential filtering.