Automating 3D Scanning of Factory Hall by a Mobile Robot

These days, factories are often planned in simulation environments – digital twins first. The advantage of these digital twins is that the preparation and testing the entire robotized production line can be performed before the actual factory hall is built. The process assumes the equipment locations in the factory correspond exactly the digital twin ones. Therefore, it is necessary to check the actual locations and fix inaccuracies either in the factory or in its simulation. This work proposes a method of autonomous 3D scanning in factory halls with a mobile robotic platform. As a main part of the method, we designed a new exploration algorithm, called Exploration by Static Scans (ESS), that optimizes the completeness of 3D data needed for the installation check in combination with minimizing the time needed for the process. Due to technical limitations, the high-resolution point clouds are not available in real-time during the factory mapping process. The algorithm estimates several scanning positions based on the digital twin model. The following scanning positions are determined realtime from low-quality LiDAR data. At first, we verified the method in a model of the factory hall in a robotic simulator. Then we assembled a hardware prototype of a scanning mobile robotic platform and verified its functionality in a real factory hall environment. Finally, we summarized the obtained knowledge of using an autonomous robot to automate 3D scanning in factory halls in a study of technical feasibility.