An architecture for passive joint localization and environment learning in shallow-water underwater acoustic settings

Abstract

Passive localization and tracking of a mobile emitter, and joint learning of its reverberant 3D environment, is a challenging task in various application domains. In underwater acoustic monitoring with a receiver array, for example, a submarine may need to be tracked in a setting with natural and man-made obstacles, such as seamounts or piers. If such obstacles occlude the line of sight from this vessel to the receivers, then the non-line of sight reflected arrivals from the reverberant environment must be leveraged for localization. Hence, we need to precisely map these reflective features in order to deliver robust performance. We propose a multi-stage global optimization and tracking architecture to approach this problem. Each stage of this architecture establishes domain knowledge such as synchronization and occluder mapping, which are inputs for the following stages of more refined algorithms. This approach is generalizable to different physical scales, and improves on methods that do not exploit emitter motion. We further introduce a robust neural network-based reflector estimation method that outperforms its alternatives in realistic application settings. The performance of this holistic approach is analyzed and its reliability is demonstrated both in simulation and in a real-life reverberant watertank, which models shallow-water acoustic environments.