3D Object-Centric Panels for Human–Robot Interaction in Multi-Level Autonomy Operations : Enhancing Robotic Operations at CERN Through Mixed-Reality Object Interaction

Sammanfattning

Robots go where humans cannot. At CERN, robotic systems allow technical work to be carried out beyond the limits of safe human access, such as in areas involving radiation or limited accessibility. However, keeping the human operator outside the task area creates a new challenge: maintaining operational awareness at a distance. This thesis addresses that challenge by exploring object-centric mixed-reality interfaces for supervised robotic operations at CERN. The study investigates how task information, robot actions, documentation, and feedback can be organised around the objects that structure an operation. The work resulted in a panel system designed and implemented within the CERN Robotics GUI (CRG). The system connects selected objects to object-specific data and creates interactive panels that can be used both in the conventional 2D interface and as world-space panels in the 3D scene. The panels were tested both in simulation and on a physical LHC mock-up, demonstrating that object selection, panel creation, feedback, and robot motion can be connected in one operational pipeline. Overall, the thesis shows that object-centric panels provide a promising way to enhance supervised teleoperation through clear and task-oriented interfaces.