5G Positioning (UL-TDOA) in industrial environments with AI
A 5G live positioning of a mobile robot can be experienced at the 5G Connect. A combination of 5G UL-TDOA with machine learning is used for this purpose. Typical features of industrial environments such as shelves are included. A high availability of a precise positioning can be achieved with classic UL-TDOA methods in open spaces, but in the vicinity of objects, walls and between shelves, effects such as shadowing, scattering and reflections influence the signal propagation. By supplementing the classic approaches with methods of machine learning, precise positioning can be achieved, especially in challenging industrial environments. Fraunhofer IIS thus allows partners to test and implement industrial applications (use cases) with high accuracy requirements for positioning, such as asset tracking or security applications. For outdoor areas, 5G positioning is expanded to include GPS/GALILEO through the use of hybrid positioning approaches. This allows global and unrestricted positioning.
Fraunhofer IIS is working in parallel on a 5G Positioning SDK for locating devices such as mobile phones, AGVs and tools in 5G campus networks based on runtime measurement.
The positioning for an intralogistics application will be shown live at the 5G Connect. A mobile robot, which typically transports small load carriers, switches between an open area and areas between shelves and is localized live on site using 5G UL-TDOA and machine learning methods. The challenges and our solution approaches for radio based positioning in challenging industrial environments are shown in this example.