Requirements for an operating system for production

© Fraunhofer IPA

With FabOS, researchers are developing an operating system for manufacturing. It aims to facilitate the collection of production data and make artificial intelligence (AI) broadly usable. They have compiled the requirements that FabOS must meet in practice in a white paper.

Machine tools operate around the clock and often have lifespans of over ten years. However, estimating the appropriate intervals for maintenance and repairs is difficult. As a result, costly failures occur repeatedly. A possible solution: additional sensors could be integrated into a machine tool. They would detect and report unusual noises indicating wear and impending downtime. Intelligent algorithms could automatically evaluate the sensor data and provide clues about which component needs to be replaced.

Unfortunately, implementing such a scenario is currently very complex. An open, distributed, real-time capable, and secure operating system that enables rapid integration of AI into manufacturing would make this easier. That is exactly what FabOS is supposed to do. The 26 project partners developing FabOS have identified the practical requirements by examining potential industrial applications, such as the machine tool mentioned above.

Four requirements FabOS must fulfill

Through discussions with industry, the researchers identified four requirements that FabOS must meet:

1. Real-time capability – Production data must be collected and evaluated simultaneously. Only then can impending disruptions be responded to before they halt operations.
2. Data security – The process knowledge of manufacturing companies must not fall into the wrong hands, as they could misuse this knowledge for their own purposes.
3. Seamless Computing – The computer resources of a company must be accessed uniformly. For example, software can be run on a suitable computer that currently has enough capacity. This allows the entire computing power of a company to be used more efficiently.
4. Industrial Edge Networks – The local company network must integrate not only IT infrastructure but also the machinery. This enables machine data to be read out and processed automatically.

The insights of the 26 project partners from research institutions, universities, and companies are documented in the English-language white paper "Industrial Application Areas and Requirements for the FabOS Platform." It is available free of charge.