Applying artificial intelligence in practice

On the way to personalized products, Artificial Intelligence (AI) provides a tremendous boost in productivity. To advance research and technology transfer of AI and machine learning (ML), scientists from the Fraunhofer Institute for Production Technology and Automation IPA have come together to form the Center for Cyber Cognitive Intelligence (ZCCI). On Open Lab Day on July 5, 2018, they will not only provide insights into their work but also offer company representatives the opportunity to implement their application ideas within the framework of a field study.

In Industry 4.0, there is a great opportunity for the economic location of Germany. Because productions of various sizes and sectors can further improve productivity, competitiveness, and quality of their work processes with intelligent automation solutions. And for manufacturers of machinery and systems, completely new innovation potentials arise. The reason: thanks to increasing networking of production facilities, vast amounts of data are available. When analyzed with AI and ML methods, on the one hand, performance leaps in the use of machines and systems in terms of quality, flexibility, and efficiency are expected. On the other hand, AI-equipped robots will increasingly enter all work and everyday environments – and take on new tasks.

AI and machine learning also play a key role in the production of personalized products at mass-market prices: "Machines and robots can hardly be retrofitted and reprogrammed for each individual personalized product. They must be able to adapt independently to changing conditions through learning algorithms," says Werner Kraus, Deputy Head of the Robotics and Assistance Systems Department at Fraunhofer IPA. How this can succeed is being researched by Kraus and scientists from other departments of Fraunhofer IPA. They have come together for this purpose in the ZCCI.

Quick Checks for Application Ideas

Insights into ongoing research projects will be provided by the scientists on Thursday, July 5, 2018, during Open Lab Day at ZCCI. In lectures and through various demonstrations, they will present a broad spectrum of machine learning methods for the

- Robotics – In the "Deep Grasping" project, for example, researchers apply ML to grasping in a box. The necessary data are collected in a virtual learning environment, and the pre-trained neural networks are then transferred to the real robot.
- Optimization of production processes – The self-learning analysis tool "Smart System Optimization" detects and reports errors and their causes in ongoing production processes independently.
- Environment sensing – In the projects "Ambos" and "MonSiKo," machine learning methods are used for gesture recognition to support manual work steps at disabled workstations or assembly stations.
- Quality assurance – Deep learning is used, for example, to evaluate laser-cut edges on sheets using image data. The evaluation criteria are roughness and burrs. Based on the results, the optimization of the cut edge quality will follow by improving the processing parameters.

Furthermore, visitors of Open Lab Day will have the opportunity to discuss their ideas for applying AI and ML in the four areas of robotics, process optimization, environment sensing, and quality assurance with the attending scientists and to apply for one of the 40 Quick Checks in total.

The winners will be allowed to realize their application idea in the context of a field study together with experts from Fraunhofer IPA. The twelve most promising application cases from the Quick Checks will later be examined in more detail in Exploring Projects. Ultimately, the experts from Fraunhofer IPA will develop six application cases into demonstrators. Because knowledge and technology transfer are the central tasks of ZCCI.

About ZCCI

The ZCCI advances the use of machine learning methods in industrial production processes through applied research by transferring knowledge and technologies into practice. It specifically supports small and medium-sized enterprises in testing and implementing customized, scalable solutions in a needs-based and precise manner. The ZCCI started its work on April 1, 2018. Together with seven other institutions, it is part of the Performance Center for Mass Personalization at the Stuttgart Technology and Innovation Campus (S-TEC).