Relocating drifting sensor networks with ambient noise correlations

Abstract

A network of drifting sensors, such as hydrophones mounted to freely drifting buoys, can be used as an array for locating acoustic sources underwater. However, for accurate localization of such a source using coherent processing, the positions of the sensors need to be known to a high degree of accuracy, typically more accurately than provided by dead reckoning or GPS alone. Past work has demonstrated the inter-sensor distances can be obtained from long-term ambient noise correlations on fixed arrays [Sabra et al., IEEE J. Ocean Engineering, 2005, 30]. Here, the approach was extended for tracking drifting sensor motion by combining a stochastic search algorithm with ambient noise correlation processing. Optimization of the stochastic search method was explored and performance compared to acoustic data collected from a volumetric hydrophone vs. vector sensor array deployed in the Long Island Sound.