Aitor Iriondo Pascual
School of Engineering Science
AIRGO is developing an AI-based tool that will make it easier for industry to identify ergonomic risks before they lead to injury. By using cameras, sensors and machine learning, workplaces can be assessed more objectively, faster and in a way that is accessible to more people. In the long term, the results may help reduce work-related strain injuries and improve working conditions, also in other sectors involving manual work.
"We want to help companies identify ergonomic risks in time and give them clear information they can use to take action. With AIRGO, we hope to contribute to fewer work-related strain injuries and a better, more sustainable working environment."
Aitor Iriondo Pascual, Project Leader and Post-Doctor in Product Design Engineering
Work-related musculoskeletal disorders are a significant problem in manufacturing. They cause suffering for workers and impose large costs on society. In 2017, the EU agency EU-OSHA estimated the annual cost across the EU at around 476 billion euros.
To prevent these injuries, companies need to assess the ergonomics of their workstations. The work is currently dependent on specialist expertise, it takes time, and the results can vary between assessors. As a consequence, many risks are identified late or not at all.
AIRGO develops a digital tool that makes ergonomics assessments more objective and more widely available. The tool uses ordinary cameras and force sensors to record how an operator moves and what loads the work places on the body. Machine-learning models identify the critical movements and select an appropriate assessment method, and a language model summarises the result in a report that operators, ergonomists and production engineers can use in their daily work.
The tool is developed together with two industrial partners that represent different parts of Swedish manufacturing. Skandia Elevator in Vara is a family-owned SMI company, and Alfa Laval in Lund is a large international manufacturer. The work is organised in three iterations called bicycle, motorcycle and car. Between iterations, the tool is evaluated during normal production at the partner sites, and the development builds on what works.
The results will be released as open source. Beyond manufacturing, the methods are likely to be useful in other sectors that involve manual work, such as healthcare and construction. AIRGO is funded by the Knowledge Foundation (KK-stiftelsen) and runs at the Virtual Engineering research environment, University of Skövde.