Pushing digital process optimization
Chemnitz University of Technology develops learning algorithm for improved car body manufacturing in BMBF-funded project
Prof. Dr. Fred Hamker © Lili Hofmann
/TU Chemnitz, Fejes, Karnitz/
The digitization of production is currently one of the most important fields of action to secure future growth and employment in Germany’s businesses in which the production of car bodies and related technologies is a central branch.
The new research project called “Machine Learning for the Prognosis of Process Parameters and Component Quality in Automotive Body Manufacturing” (ML@Karoprod) is funded by the Federal Ministry of Education and Research (BMBF). Project partners are the Professorship of Artificial Intelligence (Prof. Dr. Fred Hamker), the Fraunhofer Institute for Machine Tools and Forming Technology IWU in Dresden (Dr. Mathias Jäckel) as well as Scale GmbH (Dr. Ingolf Lepenies). The focuses on the model development and the application of machine learning (ML) techniques to accelerate the initial planning and serial setup in car body manufacturing. Chemnitz University of Technology will receive € 257 000 of the total funding of about € 1.2 millions.
The Chemnitz University of Technology schedule follows three main phases: In the first phase, a forward model is trained in order to predict the consequences of an action with unsupervised learning. In the second phase, this trained model will be used to accelerate the model-based Deep Reinforcement Learning Algorithm. Its task will be to find a sequence of production parameters that optimizes the quality of the final product. In the third phase, the developed algorithm is tested on the real production system in which its robustness against missing sensory information will be tested and evaluated.