Environmental field estimation by consensus based dynamic sensor networks and underwater gliders

Abstract

A coordinated dynamic sensor network of autonomous underwater gliders to estimate 3D time-varying environmental fields is proposed and tested. Each sensor performs local Kalman filter sequential field estimation. A network of surface relay nodes and asynchronous consensus are used to distribute local information among all nodes so that they can converge to an estimate of the global field. Tests using data from real oceanographic forecast models demonstrate the feasibility of the approach with relative error performance within 10%.