Passive detection on a distributed network using information geometry

Abstract

Using data from the recent “Distributed Network Consensus” (DSNCON19) experiment several aspects of information geometric source detection are tested. In this experiment, acoustic data were collected on seven spatially separated bottom mounted arrays in a “truss” configuration approximately 60 NM off the coast of New Jersey using towed and autonomous CW sources. Recent proposals for new detectors based on the averaging of covariances matrices from a single array according to principles of Riemannian geometry are tested against conventional passive detection schemes. In addition, new information-geometric passive detectors are developed for the distributed network of arrays. The possibility of source localization using the network is also considered. [Work supported by the Office of Naval Research.]