Every one who has already tried to install something knows that there is often a big difference between what needs to be done to just get the work done and what actually works. Within the production line of factories this means that workers often face problems. When they try to follow the Standard Operating Instructions (SOIs), those rarely represent the actual situation.
To solve this problem EIT Digital, a leading European organization for digital innovation and entrepreneurial education, started the innovation activity AI-Aid “Just-in-Time AI-Based Problem-Solving and Training Support for the Digital Factory Floor” in 2018. The project faces the problem of the growing knowledge divide between highly qualified employees and workers from assembly lines, evoked by the Internet of Things (IoT).
Taking the example of project partner Airbus, one of the biggest players within the aerospace industry, this means: the final assembly of aircrafts remains a labor-intensive process. It requires highly skilled technicians who often have to work in narrow spaces with components whose installation does not tolerate much. Due to its manual nature, the degree of complexity and the number of components, there is a risk of human error. Errors can be expensive and time consuming because a single component, that has been assembled wrongly or broken during the process, will lead to large parts of the aircraft having to be disassembled and reassembled.
The AI-Aid digital innovation activity assumes that Artificial Intelligence knowledge support systems can compensate the lack of knowledge in highly complex manufacturing halls by supporting employees individually and contextually based on their previous man-machine interaction. The goal is to provide problem-solving assistance on site, through step-by-step manuals that respond to employee actions and long-term development by providing relevant information.
DFKI leads the activity. Other partners in the initiative are Digital Catapult from London, which brings its know-how on Internet of Things connectivity solutions, as well as business champion NEOCOSMO, a German company working with modern learning environments. The Airbus Group participates as a pilot customer.
The project simulated a part of an aircraft interior containing Digital Catapult sensors for various critical components. These sensors record, for example, whether a connection was made correctly, in what order different components were installed, and whether each component was properly aligned.
DFKI and NECOSMO use the obtained data to provide a digital, adaptive knowledge and learning environment. The basis for this is the artificial intelligence-based knowledge and support service for intelligent production (“adwisar 2.0”). This architecture interprets the data of the production line to provide possible support measures for the workers.
NEOCOSMO integrated this service into the microlearning and communication platform PIIPE, which serves as a knowledge service in the project. To do this, the personalization technology in PIIPE was expanded to include intelligence support developed by DFKI to provide the opportunity of proactively identifying the individual knowledge needs of factory-based employees and to provide the knowledge needed to address individual gaps.
On PIIPE, users can find articles, dossiers, step-by-step instructions, videos and other learning materials in the sense of “learning nuggets”, regardless of time and place. A company can offer its employees content channels that they can subscribe to. Once new content such as articles, videos or knowledge dossiers have been published or updated, PIIPE notifies users and helps keeping them up-to-date.
The one-year EIT project has proved to AI-Aid that IoT assets can gather valuable data to make manufacturing processes and training more effective. The future Airbus factory has implemented a multi-network solution that enables the company to communicate in a way that was previously unachievable.