ErgoMaps: Towards Interpretable and Accessible Automated Ergonomic Analysis

Abstract

Work-related musculoskeletal disorders account for high levels of absence and deplete the workforce through permanent incapacities. Degenerative health conditions can be prevented, if ergonomic risks are discovered and eliminated before the consequences can settle in. To prevent ergonomic risks in the early stage, ergonomic analysis is necessary. However, analysis conducted manually can lead to subjective results and is demanding in terms of time and effort, while current automated techniques are not easily accessible and lack interpretability. We propose ErgoMaps, a new method for automated ergonomic risk identification that results in interpretable heatmaps. Our tool is based on human pose estimation and RULA assessment method. It takes an RGB video as input and is open-sourced to increase accessibility. Our qualitative study with eight experts shows that visualisations in the form of heatmaps can be easily adopted by users and are well suited for first level risk assessment. Furthermore, study participants believe that ErgoMaps can reduce the time needed to conduct analysis and improve the awareness of ergonomic risk evaluation among non-experts.