Estimation filtering for Deep Water Navigation

Costanzi, Riccardo; Fenucci, Davide; Caiti, Andrea; Micheli, Michele; Vermeij, Arjan; Tesei, Alessandra; Munafo, Andrea. 2018 Estimation filtering for Deep Water Navigation. IFAC-PapersOnLine, 51 (29). 299-304. https://doi.org/10.1016/j.ifacol.2018.09.519

Before downloading, please read NORA policies.

Preview

Text
1-s2.0-S2405896318322080-main.pdf
Download (866kB) | Preview

Official URL: http://dx.doi.org/10.1016/j.ifacol.2018.09.519

Abstract/Summary

The navigation task for Unmanned Underwater Vehicles is made difficult in a deep water scenario because of the lack of bottom lock for Doppler Velocity Log (DVL). This is due to the operating altitude that, for this kind of applications, is typically greater than the sensor maximum range. The effect is that the velocity measurements are biased by sea currents resulting in a rapidly increasing estimation error drift. The solution proposed in this work is based on a distributed, cooperative strategy strongly relying on an acoustic underwater network. According to the distributed philosophy, an instance of a specifically designed navigation filter (named DWNF - Deep Water Navigation Filter) is executed by each vehicle. Each DWNF relies on different Extended Kalman Filters (EKFs) running in parallel on-board: one for own navigation state estimation (AUV-EKF), the other ones for the navigation state of the remaining assets (Asset-EKF). The AUV-EKF is designed to simultaneously estimate the vehicle position and the sea current for more reliable predictions. The DWNF builds in real-time a database of past measurements and estimations; in this way it can correctly deal with delayed information. An outlier detection and rejection policy based on the Mahalanobis distance associated to each measurement is implemented. The experimental validation of the proposed approach took place in a deep water scenario during the Dynamic Mongoose’17 exercise off the South coast of Iceland (June-July 2017); preliminary analysis of the results is presented.

Item Type:	Publication - Article
Digital Object Identifier (DOI):	https://doi.org/10.1016/j.ifacol.2018.09.519
ISSN:	24058963
Date made live:	13 Nov 2018 14:33 +0 (UTC)
URI:	https://nora.nerc.ac.uk/id/eprint/521495

Item Type:

Publication - Article

Digital Object Identifier (DOI):

https://doi.org/10.1016/j.ifacol.2018.09.519

ISSN:

24058963

Date made live:

13 Nov 2018 14:33 +0 (UTC)

URI:

https://nora.nerc.ac.uk/id/eprint/521495