The 35-day trial was conducted at the United Kingdom Atomic Energy Authority’s (UKAEA) Joint European Torus (JET) facility in Oxford, which was one of the largest and most powerful fusion research machines before its recent operational shutdown.
The UKAEA, in collaboration with the Oxford Robotics Institute (ORI) at the University of Oxford, demonstrated that part of a fusion facility’s maintenance can be carried out by an autonomous robot.
This breakthrough is a step towards the commercial viability of fusion energy, making such work safer for humans and more cost-effective for its operator.
Dr Robert Skilton, head of research at UKAEA’s Remote Applications in Challenging Environments division, said: “The project aimed to validate the reliability of autonomous robotic technology and instil trust and confidence in their use for safe and efficient inspections in fusion facilities over extended periods.”
The JET facility, which remains hazardous after two high-powered deuterium-tritium experiments in the past three years, provided the ideal environment to test ORI’s autonomy platform ‘AutoInspect’.
The platform controlled a Boston Dynamics Spot, which resembles a robot dog, which was tasked with mapping the entire facility, taking sensor readings of its environment, and avoiding obstacles and personnel involved in the decommissioning process.
Dr Skilton added: “This deployment demonstrates that autonomous robots can enhance safety and cut costs.
“These ‘next generation’ solutions are becoming ready to be used in other industrial facilities such as nuclear decommissioning, environmental clean-up, and disaster relief.”
The inspection tasks were made possible by integrating UKAEA’s inspection payload with ORI’s localisation and mission autonomy solutions.
These were developed through a collaboration between Professors Hawes’ and Fallon’s teams, and combined with Boston Dynamics’ hardware and collision avoidance technology to enhance the robot’s capabilities.
The integrated robotic system collected essential data on JET’s environment and overall status twice a day, allowing the team to assess the feasibility of replacing human inspections with fully autonomous processes.
Professor Nick Hawes, professor of AI and robotics at the university, said: “Projects like this demonstrate the value of autonomous robots – robots that can do things themselves without direct control of humans.
“They also ground our science in real use cases, and provide requirements and constraints that drive us to invent new AI and robotics algorithms.”
The results of this project will assist planning for the next stages of the JET decommissioning and repurposing programme.
The ORI’s work was supported by the Engineering and Physical Sciences Research Council through the Embodied Intelligence: From Sensing to Collaboration programme grant.