Industrial robots are precise, fast and powerful. In the changed situation of the past two years, however, one specific shortcoming has become apparent: flexibility. The project "GANResilRob - Generative Adversarial Networks and Semantics for Resilient, Flexible Production Robots" presented by the Karlsruhe Research Center for Information Technology FZI therefore focuses on the anticipation of exceptional situations. In the production of industrial companies, a large number of people often work in a confined space and in closed rooms. In the event of a pandemic or political conflict, for example, this can pose a risk that endangers the health of employees and the productivity of the company. The "GANResilRob" approach creates a flexible, resilient industry based on the combination of machine learning, Generative Adversarial Networks (GAN), artificial intelligence (AI)-based semantic interpretation and intuitive task programming.

For cases in which, for example, material supply chains are interrupted, the FZI Research Center for Information Technology will develop production systems that learn tasks remotely and automatically adapt production to the changed conditions. The FZI's tasks include researching the flexible and resilient robotic assembly of vacuum chambers and possibilities for recycling components from old cars.

Video