Reconstructing JET using LIDAR-Vision fusion

Abstract Commercialising Nuclear Fusion requires considerable advances in sensors for accurately localising Remote Maintenance systems. One promising technology is LIDAR, frequently used for robotic positioning and navigation. We present work describing the 3D mapping of the inside of the Joint European Torus using a combined LIDAR-Vision fusion based measurement and navigation system from the Oxford Robotics Institute. We generate a point cloud model from our measurements and compare it with CAD models of the JET vessel using numerical methods. We show that millimeter and sub-millimeter accuracy of results are possible under the right circumstances, with nearly 10% of points having a measurement error of between 2.5 and 0.03 mm compared to the CAD model. We briefly review the potential of radiation hardening LIDAR scanners for wider use in Fusion contexts. Finally we draw conclusions about the applicability of LIDAR systems to mapping and localisation problems within fusion environments, and detail further work required.