At the RAPP Center, we conduct experiments and analyze data in problem areas that require the development of special new techniques or methods, such as distinguishing between signal and noise, complex event reconstruction, or solving problems related to condition monitoring and predictive planning. Many of these requirements also arise in similar forms in technical and industrial applications.

Some of this work is carried out in cooperation with the Lamarr Institute for Machine Learning and Artificial Intelligence. Together, we develop and test machine learning methods on real physical experiments such as the FACT telescope, the MAGIC and LST telescope systems, and the cubic-kilometer-sized IceCube neutrino telescope at the South Pole. There, methods are used under sometimes extreme operating conditions (in terms of technical requirements but also available resources), for example to quickly separate interesting physical events from background signals and to make efficient use of limited resources. One advantage of this constellation is that new methods can first be tested on physical data and setups without affecting sensitive industrial data or data rights. The methods can then be adapted to other domains, such as medical imaging.

In earlier work, instrumentation for FACT was developed, including a flat silicon photomultiplier camera. Such comparatively inexpensive and finely pixelable detectors are also of interest for applications in nuclear medicine, for example with regard to the relationship between cost, necessary dose, and achievable image quality.

Similar methodological building blocks, such as condition monitoring, separation of signals and interference, or the planning of interventions, can also be found in distributed technical systems with limited resources. We see further opportunities for transfer and joint developments in this area.
If you are interested in exchange or cooperation, please feel free to contact us.