This Project reviewed the state of the art in robotics and teleoperation inside and outside the nuclear industry to form recommendations as to how new technologies can be applied to address the significant challenges faced in the nuclear industry. Scenario-based roadmapping was undertaken, producing an assessment of key capabilities requiring nuclear-specific development, as well as an understanding of available robotics technologies that could be quickly brought across to the industry, and initial concept prototypes were produced.

In robotics, being able to plan robustly and efficiently in a complex environment is still a challenge, yet so important to make it easy to teach and program robots to be able to operate in dynamic workspaces. We will tackle this problem by benchmarking state-of-the-art planning algorithms on a list of relevant problems and use this to test the feasibility of a new approach to engaging multiple planning strategies. We will disseminate our results and developments benefiting from an Open Source approach to the community. This will speed up future development of new planning algorithms.

Robots should be used more widely to perform difficult, dirty and dangerous tasks. At present, there is some use of “remote handling” and “remote manipulation” robots in these tasks, but if the robots were more capable, it would be easier to do, and the systems would perform better. Recent research in how to get robots to manipulate objects is now being tested to see if it can be used to solve real-world problems. The GCS project will test new designs of sensing and new algorithms for manipulation to understand how they can be applied to make robots more capable of handling objects in the real world, which in the future should mean more robots, and fewer humans, in dangerous places.

We aim to create a novel semi-autonomous teleoperation experience using modern virtual reality tools. We will build on Shadow Robot’s autonomous dexterous grasping capabilities, Cambrian’s teleoperation control platform and OC Robotics flexible snake arm robots, to take advantage of significant RAS, teleoperation research and equipment at UCL to build a new capability. Our model of teleoperation control will use augmented reality and gesture recognition to drive robots, with autonomous grasping technologies used to hold objects and automatic path planning to manage motion in complex workspaces.

Smart vision for grasping robots (like the Shadow Smart Grasping System) will unlock significant new markets in research and industry. The iSee project is a feasibility investigation to find out if linking state of the art robotics hardware with cutting-edge research in vision and modern “deep learning” methods can transform the way robots can see - and therefore interact with - the world.