In the DroneStock research project, an unmanned aerial system was developed for automatic stocktaking and quality inspection of pallet contents in indoor block storage facilities. By using a helium-based lighter-than-air drone and intelligent sensor technology, the drone can navigate collision-free and does not need to be separated from humans or other automated systems. Latency-critical calculations, e.g. for collision avoidance, can be carried out directly on the drone using special AI hardware. Complex calculations for package recognition and identification (QR codes) are outsourced to a mobile server, which is positioned directly in the indoor area and has its own wireless communication system. A separate 3D environment was developed to control and visualize the system.

To the video