Spatial relationship-aware rapid entire body fuzzy assessment method for prevention of work-related musculoskeletal disorders

In the advent of Industry 5.0, advances in human-centered smart manufacturing (HSM) accentuate the role of humans in human-machine collaboration. This development has catapulted human health in human-machine systems to the forefront of the conversation. Although various tools have emerged to mitigate work-related musculoskeletal disorders (WMSDs), combining biomechanics with human morphology, the extant methods primarily hinge on expert scoring. Such methods display a step-wise change between risk levels, yielding inadequate assessment accuracy and posing challenges to human health assurance in HSM. To address these issues, this study proposes a spatial relationship-aware rapid entire body fuzzy assessment technique. The proposed method enhances the rapid entire body assessment (REBA) by enacting a dynamic evaluation of WMSDrelated risk via a deep learning-based 3D pose reconstruction. Contrary to the step-wise transitions between REBA's different risk levels, the proposed method actualizes a fuzzy assessment of WMSD risk by introducing weights between these levels. This innovation allows for a more accurate risk assessment for workers engaged in HSM. Validation through experiments conducted on data from an automobile production line demonstrates that the proposed method can achieve a precision rate of 99.31%. Demo videos and code are available at https://gith ub.com/giim-hf-lab/REBA-PLUS.

Automated summary

In the era of Industry 5.0, human-centered smart manufacturing (HSM) has emphasized the role of humans in collaboration with machines. This study proposes a method that combines deep learning-based 3D pose reconstruction with rapid entire body assessment (REBA) to assess the risk of work-related musculoskeletal disorders (WMSDs) in HSM. The proposed method improves the accuracy of risk assessment by introducing weights between different risk levels, leading to a precision rate of 99.31% in experiments conducted on an automobile production line.