Collaborative Technology Hardened for Underwater and Littoral Hazardous Environments

"QinetiQ, a UK multinational defence technology company based in Farnborough, Hampshire, has teamed up with a number of the UK's top innovative technology providers in response to Innovate UK's Research and Developments competition for Robotics and Artificial Intelligence in extreme and challenging Environments. The title of the project is ""Cthulhu"" named after the cosmic entity created by writer H. P. Lovecraft, a gigantic entity worshipped by cultists. Cthulhu's appearance is described as looking like an octopus, a dragon and a caricature of human form.

QinetiQ have assembled a comprehensive team suited for the complex and wide ranging challenges associated with the decommissioning of the active process plants on the Sellafield site. The team includes expertise from both industry and academia, and the fields of expertise include: QinetiQ (lead), Nuvia UK Ltd, University of Lancaster, Bristol Maritime Robotics and FORTIS Remote Technology ; all the partners have in the past or are currently already working with Sellafield Limited and also across the Nuclear Decommissioning Authority (NDA) estate on an number of diverse decommissioning related projects, bringing together complimentary technologies, systems, understanding and skills to deliver solutions for extreme environments that have application, some cross-cutting, Nationally and Internationally.

This project undertakes research and the development of autonomous systems that exploit state of the art machine learning technologies in order to facilitate Autonomous Inspection and Maintenance of Hazardous (nuclear) Spaces.

The proposed solution will deliver the following innovative components:

* A robust robotic platform that is amphibious, with higher level of autonomy for extreme environment operations with 24/7 availability.
* Simultaneous Localisation And Mapping (SLAM) based on sonar, tactile and passive Electro-optical (EO) sensors enabling underwater operations; able to recognise objects of interest using new fast transparent deep learning image classifiers, make decisions in the context of the task (inspect and move) including collision detection and avoidance
* Tactile sensing for visually obscure environments to enable detailed local situational awareness to be achieved in support of the sonar sensing.
* The platform will be compatible with a range of intelligent tooling modules and hence adaptable for a range of operational scenarios."