PROGRAM

Keynotes

AUV2020 will feature three plenary speakers, Mario Brito, Anna Wåhlin, and Yanwu Zhang.

Mario Brito

Dr. Mario P. Brito is an Associate Professor in Risk Analysis and Risk Management at the University of Southampton. He is a Vice-Chair of the Society for Underwater Technology Special Panel on Untethered Autonomous Systems. He is a Co-Chair in the European Safety and Reliability Association section on Maritime and Offshore Technology. He has conducted risk analysis for underwater autonomous systems deployment under ice for the Natural Resources Canada and for the Natural and Environment Research Council, for the ISE Explorer and Autosub 3 AUVs respectively. He has lead and supported accident investigations of several autonomous underwater vehicles, including Autonaut (NERC) and Nereus (Woods Hole Oceanographic Institute, Boston). Dr. Brito’s research on Autonomous Systems Risk Management has been published in journals such as Risk Analysis Journal, IEEE Transactions on Engineering Management, Journal of Atmospheric and Oceanic Technology and Antarctic Science. Dr. Brito was the PI for a Knowledge Transfer Partnership Project with ASV Ltd on Risk and Reliability Methods for Marine Autonomous systems (£166K), which has resulted in an impact case for the University of Southampton. Dr. Brito was the NERC PI for EU Horizon2020 project Bringing together Research and Industry for the Development of Glider Environmental Services (Euro1.7M), he was the Work-Package 7 leader, Systems Integration and Reliability.

ABSTRACT
Towards using Machine Learning for Autonomous Underwater Vehicles Mission Risk Quantification

Important science questions can be answered by in-situ measurements underneath sea ice and ice shelves. Over the years, autonomous underwater vehicles provided means for conducting such measurements and observations. Autonomous underwater vehicles missions in these environments present a very high risk of AUV loss. Managing the risk of AUV loss is not a trivial task. The practice of having copies of the same AUV does not increase the chance of success as systemic failures may occur when the same environmental conditions are observed. On the other hand, risk models developed for quantifying the risk of AUV loss based on technical failures rely heavily on expert judgments and this presents limitations introduced by potential motivational and subconscious bias.

Therefore validation of risk models is essential for increasing trust in AUVs missions in extreme environments. In this talk I present methods for validating experts’ assessments on the risk of AUV loss and experts’ assessments on the effectiveness of risk mitigation actions. I discuss how this can be used for estimating the reliability of the risk model. The results of the validation are used for explaining the fallacies in expert judgment elicitation and to make recommendations for improvements.

The talk then presents a method based on machine learning to estimate the risk of AUV loss. Risk updating methods to augment the proposed machine learning approach for risk assessment are discussed.

Anna Wåhlin

Anna Wåhlin is a Professor of Physical Oceanography at the Department of Marine Sciences, University of Gothenburg. Her research focus is in the field of Polar Oceanography, mostly in the Southern Ocean. Specifically, her research investigates several aspects of dynamics of polar seas, including physical oceanography, ocean circulation, topographic effects, ice shelf melt processes and air-sea-ice interaction. When Wåhlin was appointed professor in 2015, she became Sweden’s first female full Professor of Oceanography. She is project leader for Sweden’s national AUV infrastructure funded by the Knut and Alice Wallenberg Foundation. This AUV became the world’s first to venture under Thwaites glacier, Antarctica, in 2019. Between 2015 and 2017, Wåhlin was co-chair of the joint Scientific Committee on Antarctic Research (SCAR) and SCOR initiative Southern Ocean Observing System (SOOS). She is an Associate Editor of the journal Advances in Polar Science and member of the IOW scientific advisory board (2016-2019). Her awards include being a Fulbright Scholar (2007-2008), receiving a Crafoord Research Stipend from the Swedish Royal Academy of Science (2010), being a SCAR visiting professor (2013) and receiving the Albert Wallin science prize from the Royal Society of Arts and Sciences (KKVS) in Gothenburg 2018.

ABSTRACT
An AUV underneath the ‘Doomsday glacier’: Revealing pathways and modification of warm water flowing beneath Thwaites ice shelf, West Antarctica

The Swedish AUV Ran, a Kongsberg Hugin AUV with 3000 m depth rating, was operated during the N.B. Palmer expedition to the Thwaites glacier (Antarctica) in February and March 2019. In this presentation, the main challenges and lessons learned will be presented together with some of the key findings from the expedition. The performance of the AUV will be discussed as well as the value of the various payload.

The fate of the West Antarctic Ice Sheet is the largest remaining uncertainty in predicting sea-level rise through the next century, and its most vulnerable and rapidly changing outlet is Thwaites Glacier. Because the seabed slope under the glacier is retrograde (downhill inland), ice discharge from Thwaites Glacier is potentially unstable to melting of the underside of its floating ice shelf and grounding line retreat, both of which can be enhanced by warm ocean water circulating underneath the ice shelf. Here we present the results of two missions underneath Thwaites ice shelf performed by the AUV ‘Ran’: The very first direct observations of ocean temperature, salinity, and oxygen underneath Thwaites ice shelf. Using the high precision environment payload suite, observations were obtained that indicate deep water (> 800 m) underneath the central part of the ice shelf is in connection with Pine Island Bay, a previously unknown westward branch of warm deep water entering the ice shelf cavity. Warm water also enters from the north in two troughs separated by a pinning point. Spatial gradients of salinity, temperature and oxygen recorded underneath the ice shelf indicate that this is an active region where several water masses meet and mixes. The observations identify the central buttressing point as a vulnerable region of change currently under attack by warm water inflow from all sides: a scenario that may lead to ungrounding and retreat more quickly than previously expected.

Yanwu Zhang

Yanwu Zhang is a Senior Research Engineer at the Monterey Bay Aquarium Research Institute (MBARI). He received the Ph.D. degree in oceanographic engineering from the MIT/WHOI Joint Program in 2000. At MBARI, he leads the project of targeted sampling by autonomous vehicles, designs adaptive sampling algorithms for marine ecosystem studies, and participates in the development of the Tethys-class long range AUVs. Since 1996, he has participated in field experiments using the Odyssey IIB, Dorado, and Tethys AUVs for a variety of research objectives, from Haro Strait tidal front mapping to microbial process studies in Hawaiian eddies.

ABSTRACT
Targeted Sampling by Autonomous Vehicles

In the vast ocean, many ecologically important phenomena are temporally episodic, localized in space, and move according to local currents. To effectively study these complex and evolving phenomena, methods that enable autonomous platforms to detect and respond to targeted phenomena are required. Such capabilities allow for directed sensing and water sample acquisition in the most relevant and informative locations. To meet this need, ocean engineers work with scientists to integrate advanced sampling devices into autonomous vehicles, and design onboard algorithms for detecting oceanic features in real time and directing vehicle and sampling behaviors as dictated by research objectives. These methods have successfully been applied in a series of field programs to study a range of phenomena such as harmful algal blooms, coastal upwelling fronts, and microbial processes in open-ocean eddies. In this talk, we review these applications and discuss future directions.