The MADEin4 project, led by iTEST GmbH with project manager Sascha Ziesig and technical lead Dr. David Buch, is a 36‑month initiative funded by the European Centre for Secure and Trustworthy Cyberspace (ECSEL JU) and the German Federal Ministry of Education and Research (BMBF). Its goal is to raise productivity in both semiconductor and automotive manufacturing by creating cyber‑physical systems that combine measurement‑data analysis, machine‑learning (ML) techniques, and digital twins. AVL DiTEST participates in three work packages: WP 1 (project management), WP 4 (a pilot line for predictive yield and tool performance), and WP 5 (digitalisation of manufacturing for higher productivity).
In WP 4 the team built a cloud‑based data platform that ingests raw on‑board diagnostic (OBD) data from vehicle fleets and from multi‑brand repair shops. The platform is implemented on Amazon Web Services (AWS) using IoT Core for data ingestion, S3 for raw‑data storage, Lambda for validation and metadata extraction, SNS for event notification, Kinesis Data Analytics for real‑time processing, Redshift for data transformation, and Elasticsearch for search and visualization. The use of the AWS Cloud Development Kit (CDK) allows rapid deployment and scaling of the analytics pipeline. By storing data in its native format the platform can adapt to future schema changes without costly migrations.
The analytical workflow correlates OBD data with end‑of‑line motor‑test‑bench results from the automotive production line. Diagnostic trouble codes (DTCs) captured in the vehicle’s control units are matched against performance metrics measured on the test bench. This correlation provides a long‑term view of vehicle reliability and performance, enabling a feed‑forward loop that transforms the manufacturing process from reactive to predictive. The platform supports the extraction of key performance indicators, the identification of root causes for failures, and the generation of actionable insights that can be fed back into the production control system.
WP 5 extends this approach to the broader manufacturing environment, integrating the predictive analytics pipeline with plant‑level control systems. The result is a digital twin of the production line that continuously learns from real‑time data, allowing operators to anticipate tool wear, adjust process parameters, and optimise yield before defects occur. Although specific numerical performance metrics are not reported in the interim report, the architecture is designed to handle high‑volume, high‑velocity data streams typical of automotive and semiconductor plants, and the use of AWS services ensures elastic scalability.
The collaboration structure places iTEST GmbH at the technical core, with AVL DiTEST providing domain expertise in automotive diagnostics and manufacturing. The project’s management (WP 1) coordinates activities across all partners, ensuring alignment with the 36‑month schedule and budget. The partnership leverages the complementary strengths of both companies to deliver a scalable, data‑driven solution that promises reduced production costs, shortened time‑to‑market, and improved product reliability across two high‑technology sectors.