AuQuA – Augmented Intelligence based Quality Assurance of Assembly Tasks in Global Value Networks

Key Info

Basic Information

01.04.2020 to 31.03.2022
Organizational Unit:
Chair of Production Metrology and Quality Management, Organizational Development
AiF e.V., Federal Ministry of Economics and Energy BMWi

Research partner

    • University of São Paulo
    • University of Brasília
    • denkwerk GmbH
    • Polarstern Education UG
    • pixolus GmbH
    • AC GmbH
    • sunzinet AG
    • i2solutions GmbH
    • MENNEKES Elektrotechnik GmbH & Co. KG
    • Lenord, Bauer & Co. GmbH
    • Ph-MECHANIK GmbH & Co. KG
    • Omega7 Systems Informatica LTDA
    • MM Optics LTDA
    • Exxys LTDA, Exxomed LTDS-ME
    • Bio-Art LTDA
    • Subras Moldes e Plásticos LTDA
    • Dipl.-Ing. Herwarth Reich GmbH

In this project, an assembly support system (ASS) is being developed that automatically creates, uses and continuously improves Augmented Reality (AR)-based assembly instructions using methods of Artificial Intelligence (AI). These assembly instructions serve on the one hand to guide employees in ongoing processes and on the other hand to train them in new processes. The assembly instructions are projected in real time both on the work surface and on the component to be assembled, so that workers are free of wearables.

In the first step, the system is fed with CAD data to create the assembly instructions. An assembly specialist then demonstratively assembles the product or sub-assembly. Equipped with various image acquisition sensors, the ASS recognizes the components as well as the assembly sequence and then creates the first version of the AR-based assembly instructions. The instructions can be optimized by further assembly sequences carried out by the specialist or used immediately in production.

During production, the system uses image acquisition sensors to record the current status of the assembly in real-time and checks the quality of the assembly object. Combined with real-time feedback for immediate error correction, rework and thus quality costs are reduced. In addition, the ASS can use recorded error data to continuously optimize assembly instructions.

The Human-AI-Interface is designed cognitive-ergonomically with the help of User-Centered Design (UCD), thus ensuring optimal support for humans in production. The ASS also considers the UCD in the presentation of assembly instructions and adapts to the needs of the user. This enables high flexibility in the use of the ASS in global value networks while taking cultural conditions into account.