Use Cases und Einsatzfähigkeit von Inspektionsrobotern im Bergbau

With technological advances in areas like robotics, information processing and machine learning, the performance of robots is improving constantly, leading to increasing attractiveness for applications in research and industry. Therefore, the main objectives of this work are on the one hand to gather possible use cases for robots in mining based on a literature review, and on the other hand, to examine the current usability and operability of the for the thesis available ANYmal D robot in mining environments. The latter is based on three experiments in real environments to test the practical suitability besides some smaller laboratory tests. As the third focus, the mapping and self-implemented methodology for determining a tunnel volume are put into practice as an already feasible use case for the existing robot. The results show that ANYmal D has a stable walk, except on steep or slippery slopes, and appears to be suitable for mining applications in principle. Manual measurements like mapping with the LiDAR sensor and examination of the geometry and tunnel volume can already be implemented well, although localization does not yet work error-free. Inspection tasks are by now possible, but complex and practical applications often require the demanding implementation of data interpretation methods, e.g., for the inspection of conveyor belts. For unsecured tunnel areas or rescue missions, the short radio range and lack of autonomy during exploration still pose major hurdles for practical applicability. Hence, the findings already developed by research teams for autonomous exploration should be included in the software for the robot and refined in order to make them available to smaller research teams and a wide range of industries.