Imagine a self-organizing factory: safety zones are cleared automatically and cyberattacks have no effect remain ineffective. This may sound like science fiction, but the KIbSIS research project is making it come true. As part of an international cooperation with South Korea, partners from the areas of research and business are seeking to develop fully automated AI-based safety acknowledgment for machine tools by 2028. The approach is based on standardized, trustworthy and quantum-secureafe AI technology. 

Industrial processes are becoming more and more digitalized and networked. This gives rise to opportunities as well as risks for companies. Opportunities include self-organizing factories, while risks are posed, for example, by a growing number of cyberthreats, classic cryptography-breaking quantum computers and insufficient security mechanisms for AI-based industrial systems.

The research project AI-Based Safety for Industrial Systems (KI-basierte Sicherheit für Industrie-Systeme (KIbSIS)) aims to develop technologies that help exploit the opportunities afforded by the increase increasingin digitalization and networking and to minimize related risks. These technologies include AI-based anomaly detection in the event of cyberattacks, a hardened system platform for safe AI applications, quantum-resistant cryptography using secure elements* and a trustworthy software bill of materials (SBOM)* for industrial products.

Seamless connection of cloud and edge computing

Standardized interfaces are being created as part of KibSIS, such as the Generic Trust Anchor API*, as well as quantum-resistant protocols that are compatible with various cloud-edge environments regardless of the manufacturer. KIbSIS is actively involved in IPCEI-CIS Workstream 4*, aiming to harmonize technical standards and to promote interoperable solutions. AI-based anomaly detection and automated security functions for edge devices allow enables the for systems to be scaled and made real-time capable. KIbSIS technologies therefore help to interlink cloud and edge computing to allow for a flexible distribution of resources between a central cloud and decentralized edge nodes, which in turn promotes establishment of a Multi-Provider Cloud-Edge Continuum*.

Creation of a demonstrator through international cooperationcollaboration

Combining these technologies allows for fully automated acknowledgment of the safety zone of a machine tool by means of an AI-based process, after production has been interrupted through triggering of a safety device, e.g. due to items or people tripping a light barrier. As part of this, a demonstrator is being created in cooperation collaboration with a partner from South Korea. Planning provides for the demonstrator to be presented in 2028.

Philipp Zieris, project manager at the Fraunhofer Institute for Applied and Integrated Security AISEC says: »To be able to work cost-efficiently, machine and factory operators aim to realize self-organizing production. To achieve thisthis end, automated systems need to make safety-critical decisions in a reliable and trustworthy manner. KIbSIS is dedicated to advancing technologies necessary for thisthe required technologies. By developing autonomous AI-based and quantum-resistant safety and security solutions for industrial systems, we are contributing to Germany’s and Europe’s digital sovereignty.«