Look Ma, No Hands!

A Personal Tour of Emerging Technology, Challenges and Applications for Unmanned Underwater Vehicles

David Lane

Professor of Ocean and Systems Engineering

Heriot-Watt University

Edinburgh, Scotland, UK

http://www.ece.eps.hw.ac.uk/oceans/

CEO

SeeByte Ltd

Edinburgh, Scotland,UK

http://www.seebyte.com

As we move into the 21^st century, autonomous underwater vehicles are a commercial reality, changing the way people work in numerous mapping and inspection tasks. Reliable vehicles are available that repeatably do what they are told by non-technical operators working in some of the harshest environments on the planet.

However, although they are increasingly being equipped with ever better, smaller and sometimes even cheaper sensors, operational systems still have a pretty low IQ. They don't react well to (or even notice) the unexpected, they don't make many truly independent decision, if they do then they don't brief the operator very well, they don't team easily, they still drive into things by mistake, and they still occasionally go AWOL.

Increasingly complex mission opportunities with longer deployment times in deeper water, with more loitering and following at longer ranges with less available a priori information are driving applications towards systems with greater autonomy. Research and development has responded by considering sensors, processing, architectures and algorithms that can provide improved awareness for the operator and vehicle alike, and deliver reliable autonomy.

The presentation will take a whistlestop tour through some ongoing R&D and recent demonstrations addressing this enhanced autonomy and awareness. It will also demonstrate how subsets of these technology developments and integrations are starting to be made available for use on ROVs (so called “Smart ROVs”), to alleviate lack of available skilled pilots, cost of training, volume of work, penalties for delays and the challenge of tasks being undertaken.

Examples will include a selection of topics such as :

-          real-time adaptive mission planning,  

-          co-operative multi-vehicle plan execution

-          operator information exchange through interfaces that enhance trust (through augmented reality, prediction, speech interfaces)

-          data fusion, computer aided detection and classification of objects

-          object tracking in noisy data

-          object relative motion control,

-          ontologies (data structures) for world modelling

-          smart diagnostics that detect failures before they occur, alert the operator/vehicle mission planner and are linked to maintenance scheduling

-          distributed feature based navigation (SLAM navigation) and integration with conventional navigation sensors

-          wideband multi-chirp sonar sensors,

-          synthetic aperture sonar sensors

-          RF underwater communications