Practical Experience towards Robust Underwater Navigation

Autonomous underwater operations require robust and accurate navigation capabilities, a basic requirement for both precise position and velocity estimation, as well as motion compensation and geo-referencing of exteroceptive sensor data. A great number of theoretical approaches to sensor fusion and filtering can be found in literature, but they have to be matched with practical issues such as multi-rate sampling, measurement glitching, environmental conditions. For these reasons, the preparation, tuning and exploitation of navigation systems require different development steps to achieve reliability and robustness. This paper reports the practical experience of the exploitation of a multi-module extended Kalman filter based navigation system employed on the e-URoPe AUV/ROV. The paper shows the results of motion estimation including linear velocity and position (both horizontal and vertical), as well as angular position and rate. The proposed navigation system relies on the measurement fusion gathered by a variety of sensors: USBL, DVL, FOG, AHRS, depth-meter, GPS (when surfaced). The paper provides experimental proof and ground-truth of the proposed framework. (C) 2018, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.