Artificial intelligence (AI) can already be used sensibly in production today: to monitor processes and make them more efficient, to take over routine tasks and support decision-making. But up to now there is no infrastructure that bundles AI systems and enables real-time, secure operation of data-driven services and AI applications.
Together with project partners from industry and research, the Fraunhofer IPT is developing a platform for these services and applications in order to make Industry 4.0 secure and manageable in everyday production, and to create a basis for a real-time capable cross-domain network.
This platform is designed as an open system in order to keep it as flexible as possible and to be able to develop it further. The openness of the system creates access to innovations and resources that individual companies can hardly provide on their own. The deliberate renunciation of a supercomputer in favor of a distributed system ensures scalability and reliability.
Research and industrial companies jointly define the requirements that are placed on such a platform in terms of production processes and process chains as well as the connection to machines and sensors. Based on these specifications, the platform is then built. Data standards help to reduce the effort required for the integration of AI solutions by bundling data from different connected systems. In addition, the Fraunhofer IPT is developing modules with which recorded machine data can be visualized and evaluated in the context of other data.
The individual functionalities of the platform will be tested using various application scenarios and integrated into machine tools, robotics applications and production control systems. The aim is, for example, the online visualization of product quality, anomaly detection or condition monitoring. The data management and control of machines and processes are directly connected to the platform so that all other systems can access and react to events reliably and in real time. The production systems' access to AI technologies, the bundling of data and the networking of various services and systems enables intelligent automation of machines and processes.
Development and implementation
Fraunhofer IPA, Stuttgart