Localizing Autonomous Underwater Vehicles: Experimental Evaluation of a Long Baseline Method

This work concerns underwater networking with mobile assets, like Autonomous Underwater Vehicles (AUVs), for advanced monitoring and exploration of submerged environments. Particularly, we are interested in enabling an AUV to localize itself while moving underwater by acoustically polling beacon nodes statically deployed at well-known location. Our method only relies on a model of the AUV dynamics, on an on-board depth sensor and on long baseline ranging information. The AUV applies an Extended Kalman Filter to estimate its position, without needing any further local measurements but those of depths. We have evaluated the accuracy of the proposed method via experiments at sea in the shallow waters around the Italian island of Ponza, computing the average distance between the estimated locations of the AUV and its positions as measured by GPS along its trajectory (localization error). In deployments with up to four beacons, our simple method enables AUVs to swiftly self localize with errors never exceeding 3.62m (using only two beacons), 2.65m (three beacons) and 2.45m (four beacons).