Robotics and Machine Learning

Manufacturing robot systems are characterized by a high degree of complexity that is difficult to manage, particularly by small-to-medium enterprises. For these companies the major challenge involves improving the usability, flexibility and integration of these systems. Adaptation to changing requirements should be possible without the need for expert know-how from the field of automation.

The Robotics and Machine Learning research division is integrating artificial intelligence (AI) methods and developing new concepts for collaboration between humans and robots. The aim is to derive solutions from formal product and process descriptions such that they can be flexibly ported over to different systems. Robots should be in a position to solve abstract problems in different domains on their own.

By synthesizing robot programs on the basis of declarative target descriptions, automation experts are able to specify production targets at a higher abstraction level and in a familiar language. Using formal techniques from the area of knowledge modeling, the automation and application domains are semantically (meaningfully) described in a machine-interpretable format. Gaps and ambiguities that appear in these potentially underspecified instructions are resolved with logical inferences and planning. An executable robot program can be automatically generated/synthesized as a result.

To enable the flexible reconfiguration of heterogeneous cyber-physical production systems on the basis of task specifications, we conduct research into semantic interoperability methods. This includes the automatic reconfiguration of software and hardware components via semantic resource models, as well as the comparison of semantic modeled capabilities with formal requirements, which are derived from the manufacturing processes and corresponding products.

Your contact

Dr. Markus Rickert

+49 89 3603522 43
rickert@fortiss.org