Automated Fault Diagnosis for an Autonomous Underwater Vehicle

Richard Dearden, Juhan-Peep Ernits

Research output: Contribution to journalArticlepeer-review

35 Citations (Scopus)

Abstract

This paper reports our results in using a discrete
fault diagnosis system, Livingstone 2 (L2), on-board an autonomous
underwater vehicle (AUV), Autosub 6000. Due to the
difficulty of communicating between an AUV and its operators,
AUVs can benefit particularly from increased autonomy, of which
fault diagnosis is a part. However, they are also restricted in their
power consumption. We show that a discrete diagnosis system
can detect and identify a number of faults that would threaten
the health of an AUV, while also being sufficiently lightweight
computationally to be deployed on-board the vehicle. Since AUVs
also often have their missions designed just before deployment
in response to data from previous missions, a diagnosis system
that monitors the software as well as the hardware of the system
is also very useful. We show how a software diagnosis model can
be built automatically that can be integrated with the hardware
model to diagnose the complete system. We show empirically that
on Autosub 6000 this allows us to diagnose real vehicle faults that
could potentially lead to the loss of the vehicle.
Original languageEnglish
Pages (from-to)484-499
JournalIEEE Journal of Oceanic Engineering
Volume38
Issue number3
Early online date29 Jan 2013
DOIs
Publication statusPublished - 10 Jul 2013

Fingerprint

Dive into the research topics of 'Automated Fault Diagnosis for an Autonomous Underwater Vehicle'. Together they form a unique fingerprint.

Cite this