AuQuA — Augmented Intelligence based Quality Assurance of Assembly Tasks in Global Value Networks
An international research project between Germany and Brazil to enable augmented intelligence based quality assurance of assembly tasks in global value networks.
Current challenges for the industry of tomorrow
The concept of Industry 4.0 is based on the integration of new technologies into the production of the future and it will change conventional production approaches fundamentally. One of the biggest changes in this context will be the continuing globalization of manufacturing companies. To increase competitiveness, local production sites are already being placed in emerging markets. This enables products and production to be adapted to location-specific requirements and thus to satisfy the corresponding local customer needs. To illustrate this, consider the automotive industry. While car manufacturers build mostly small and economical cars for the narrow streets and high gasoline prices of the European market, they produce mainly roomy SUVs with large engines for the spacious traffic area of the American market.
Nonetheless, with global re-allocation, not only new opportunities but also new challenges arise. On the one hand, the expansion of global value networks increases their complexity. Managing that kind of complexity requires product and production information to be available worldwide at any time. On the other hand, it is necessary to ensure a consistently high product quality level despite country-specific employee qualification levels in manual assembly. On top of that production still has to be effective and most importantly cost-efficient. This raises the big question: HOW?
- How can we assure both, a high-level of manual assembly quality and efficiency in global value networks?
- How can low-skilled workers be trained for and guided through assembly tasks?
- How can in-process quality assurance be guaranteed in manual assembly processes?
To tackle all these production-shattering questions, we just launched our international research project “Augmented Intelligence based Quality Assurance of Assembly Tasks in Global Value Networks” (AuQuA).
The Project
As an answer to these questions, we propose an assembly support system (ASS) that uses a multi-camera Context and Action Recognition System (CARS) based on methods of weak artificial intelligence, to automatically recognize manual assembly steps and create assembly instructions. These assembly instructions will then be projected with Augmented Reality (AR) technologies onto parts and / or work surfaces, to guide workers within the assembly. In simple terms, the system learns from an expert how to assemble a product and can then guide other workers to assemble the product as well. Equipped with the CARS, the ASS tracks the work progress and checks against deviations of specified processes. Assembly workers receive real-time feedback on their progress so that quality issues can be eliminated right away. This way the System generates quality data, which can be used by the system itself to automatically improve the assembly process if necessary. First tests according hand recognition and AR-based assembly instructions have already been carried out in the AIXLAB at the Machine Tool Laboratory WZL of RWTH Aachen University (see figures below).
As the assembly instructions are data-based, they can easily be distributed around the globe in seconds. This way the instructions can effortlessly be used in all production facilities of a global value network. Additionally, the ASS learns cultural differences between production locations and automatically adapts the assembly instruction accordingly (for an easier understanding illustrated in figure below). To reduce disturbance by the system, it also detects the skill level of a worker and adjusts the amount of AR-based assembly instructions to the workers’ needs. This way more skilled workers will receive less support during the assembly.
All this futuristic process automation will be based on algorithms and artificial intelligence, which we’ll develop in the next upcoming months — bringing process mining into the physical domain, so to speak. Nonetheless, the main aspect of our system ist the interaction with a human being. Without that interaction the system doesn’t work. This makes it an Augmented Intelligence application, enhancing the abilities of the worker.
To dig deeper into the technical details, read Louis’ blog post (upcoming next here).
To develop the ASS a mixed-method approach, containing usability research, social science research, as well as methods of engineering and data science research, will be applied. Just to give a brief outline of the variety of methods that will be used, here are some examples:
- Scrum
- DILO-Method
- Process Mapping
- Multi-Moment Method
- User-Centered Design
- User-Testing
- Technical development frameworks and visual prototyping
And much more.
The Outcome
The use of the ASS shortens ramp-up and work preparation time due to automatically generated continuously improved assembly instructions, guarantees quality assurance in assembly processes and delivers insights for Design-for-Assembly engineering. Consequently, the economic advantage of our research project comprises the reduction of quality costs by in-process quality assurance, reduction of assembly training costs as well as assembly time through guided assembly. On top of that, due to an easy sharing of assembly workflows as well as data-driven insights along global value networks, it helps to manage the complexity of those networks.
The Network
The research on this project and thus the development of the ASS will be conducted by an international research consortium, consisting of the Laboratory for Machine Tools and Production Engineering (WZL) of the RWTH Aachen University and the Mechatronics Group of IFSC-USP of the University of São Paulo. Both research partners will analyze social and cultural demands on the system and identify differences. Special requirements for the system will then be defined so that it can be used in both countries.
Furthermore, our consortium is enriched by numerous partners throughout different branches from the German and Brazilian industry. This way we build a comprehensive Global-Research-Network with a huge amount of knowledge and lots of different opinions. If you’re interested in the companies that we’re working with, you can view the entire list of partners on the project page.
Summary
The ASS will help to reduce quality and ramp-up costs as well as increase the overall process quality level of an assembly. Furthermore the ASS will support companies in international cooperation to build extensive global value networks. With the help of our international research consortium, consisting of German and Brazilian researchers and industry partners, we are able to focus on more than operational assembly support. We face the challenges of global collaboration, not only from a technical but from a socio-technical point of view. The data-driven insights we are going to generate will open up new possibilities for global collaboration and new business models.
Thanks for reading my article! To keep up with the whole development process, follow our Research Group.
