Bolstering the competitiveness of the machine tool industry
MINERVA research project: Privacy-enhancing technologies protect sensitive machine data
Machine learning models for predictive condition monitoring
In the “Secure collaborative utilization of machine tool data using privacy-enhancing technologies” (MINERVA) research project, the partners are using privacy-enhancing technologies to ensure machine operators’ sovereignty over the machine data, thereby protecting their intellectual property. Using anonymized data, the partners train suitable machine learning models for the predictive condition monitoring of machine tools which then work without disclosing sensitive data.
“For the machine tool industry, we are demonstrating at an interdisciplinary level that by selecting and using suitable technologies to promote data protection, the opportunities associated with sharing data can significantly outweigh the risks, not least because we are incorporating an additional level of security,” explains Bartol Filipovic, MINERVA project manager and head of the Product Protection and Industrial Security department at Fraunhofer AISEC.
Protection of sensitive data
In relation to the protection of personal data, privacy-enhancing technologies have previously been used to enable data such as patient information to be evaluated anonymously (e.g., using differential privacy). The project partners are now transferring the principles of this approach to data from the machine tool industry in order to protect the intellectual property of those disclosing the data. “If machine learning algorithms are trained with cross-company data in the cloud, it could be possible to track what kind of workpiece was manufactured (product geometry), to what extent the machine tool was utilized, how much energy was used during manufacture and the production rate used. From a user’s perspective, this is all highly sensitive information, so, in MINERVA, we are developing solutions to protect this,” says Bartol Filipovic, describing the added value of the research approach. In addition to differential privacy, the other privacy-enhancing technologies used in MINERVA include trusted execution environments, which establish a protected, attested area in the cloud, and federated learning, in which — instead of the data itself — only the ML models already trained in the edge are migrated to the cloud.
MINERVA is funded by the German Federal Ministry of Education and Research (BMBF) as part of the “IoT security in smart home, production and sensitive infrastructures” announcement. The project volume amounts to EUR 2.47 million, of which 70% is provided by the BMBF. The project will run from May 2023 until April 2026.