Artificial intelligence for automated production systems

Notwithstanding the monitoring of individual production steps, modern production systems encompass complex dependencies between materials, operating equipment, and processing steps that do not meet the required level of quality, even with supposedly correct processing. This results in costly rework or even rejects.


The challenge

The goal of the XAPS project is to use machine learning methods to identify complex dependencies between errors and processing early on, and to explain these with AI methods so the operator can efficiently and effectively optimize their automated production system. In order to accomplish this, XAPS links digital descriptions of the factory and the product, the digital twin, from the manufacturing execution system with machine learning and innovative methods of formal argumentation. The XAPS platform therefore provides the foundation for an integrated portfolio of solutions to control and monitor automated production systems.


Our contribution

Fraunhofer ISST is assisting the project consortium with collecting the requirements for the XAPS platform and preparing a software architecture. This in particular includes interfaces to collect shop floor data and transfer them to modeling.



The XAPS platform provides the user with reactive or proactive, intuitively usable explanations for problems in production. Existing configuration, control, monitoring, and sensor data of the manufacturing execution system are used by the XAPS platform to represent the physical factory as a digital factory, relating this to the digital twins of the products.



  • Old World Computing GmbH
  • iTAC Software AG
  • HELLA GmbH & Co. KGaA
  • University of Koblenz-Landau


  • Subsidized by: Federal Ministry of Education and Research (BMBF)
  • Term: 01/2020-12/2021