Academia
Remote Inspection
Platform Development – We are working to design bespoke robotic platforms to tackle various challenges including access restrictions and radiation resistance. This research will explore uses on land, in the air, and in water.
Sensor and Perception – Complex nuclear decommissioning challenges involve using different sensors to gather data, which when processed will help with decision making. We will work on both new and existing ways to detect radiation and find methods to ensure that their capabilities can be tested outside of real, active environments. We will also incorporate other sensors, like ultrasound devices, thermal cameras and devices that measure salt levels, for asset management. Machine learning techniques will process and analyse the vast data collected from these robotic operations to pull out useful information.
Mission Planning – We will develop methods that allow robots to perform tasks regularly over long periods of time. For example, they might automatically check multiple floors or spot problems in a storage facility. This will involve using advanced planning techniques, tools for making decisions based on risk, and systems that automatically detect changes.
Remote Handling
Manipulation in constrained environments – We will develop methods to help robotic arms work efficiently and safely in tight and cluttered spaces. This involves creating systems that can update maps of the area in real time, recognise objects, and automatically control the arm’s movements to avoid collisions. With a focus on remote glovebox operations, we will develop automated tools that enable point cloud images of the environment to be updated continuously, objects to be identified, and safe, low-level manipulation of the arm to be automated.
Autonomous executions of tasks – We will develop and test methods to automate simple tasks like collecting samples, opening and closing ports, and grabbing objects using the RoBox platform. We will use various techniques, including machine learning, to make this possible.
More information about RoBox can be found on our deployments page.
Human-Robot Interaction
Trust and human factor design – To help humans and robots work together effectively, with humans overseeing the work, we will analyse how to automate routine tasks using human-factor analysis. The aim is to reduce the workload of the person remotely controlling the operations (tele-operator) by determining which tasks can be automated. The results of this work will be applied to improve remote handling processes.
Reliable visual, audio and tactile interfaces – We are working to enhance the RoBox demonstrator by adding virtual reality and tactile feedback to the operator who is using tele-operations. They will improve visual and touch feedback to make the interfaces clearer and more intuitive, showing essential task details and safety limits clearly. This will help reduce the amount of mental effort required by the operator.
Verification, Safety Case Identification and Standardisation
Reliability – We will analyse and propose robotic architectures that allow for the implementation of strong verification methods.
Deployability – In collaboration with regulators and end-users, we will identify simpler processes to use robotic and AI systems in nuclear environments. We will also explore how AI can help develop the necessary safety cases to allow the use of these technologies on nuclear sites.
Resilience – We are exploring how to make robotic systems more reliable and less likely to fail during operation. We will look into various factors that can enhance the durability and dependability of these systems.
