Robin Ströbel, M.Sc.

  • 76131 Karlsruhe
    Kaiserstraße 12

Robin Ströbel, M.Sc.

Forschungs- und Arbeitsgebiete:

  • Software-defined Manufacturing und digitale Zwillinge
  • Themenbereiche der Industrie 4.0

Allgemeine Aufgaben:

Projekte:

  • SDM4FZI: Software-defined Manufacturing in der Fahrzeug- und Zulieferindustrie
  • SDMflex: Flexible SDM through Continuously Self-Learning Quality-Aware Digital Twins

Lebenslauf:

seit 04/2022

Wissenschaftlicher Mitarbeiter am Institut für Produktionstechnik (wbk) des Karlsruher Instituts für Technologie (KIT) 

10/2019-03/2022 

Studium des Maschinenbaus (M.Sc.) am Karlsruher Institut für Technologie (KIT)

09/2017-01/2018 

Auslandsstudium an der Jiaotong-Universität Shanghai (SJTU)

10/2015-09/2019 

Studium des Maschinenbaus (B.Sc.) am Karlsruher Institut für Technologie (KIT)

Veröffentlichungen

[ 1 ] Gönnheimer, P.; Ströbel, R.; Netzer, M. & Fleischer, J. (2022), „Generation of identifiable CNC reference runs with high information content for machine learning and analytic approaches to parameter identification“. Procedia CIRP, Elsevier, S. 734-739. 10.1016/j.procir.2022.05.054
Abstract
As a result of the change to Industry 4.0, the requirements for information models and digital twins are steadily increasing. Thus, reliable methods to identification and assignment of data sources in CNC machines are required. AI-based approaches are already capable of identifying individual signal groups but are increasingly reaching their limits due to the small size of existing datasets. Furthermore, the low information content of the timeseries used to build the learning datasets represents an additional limitation. In this paper, an approach is presented and examined by means of which identifiable CNC reference runs with particularly high information content can be generated to create a suitable database for machine learning approaches. Moreover, due to the uniqueness of the generated trajectories, the reference runs represent a particularly suitable basis for analytical methods to parameter identification.