Investigating the effects of back-support exoskeletons on work posture: An automated AI-based NIOSH measurement using YOLO

Work-related musculoskeletal disorders (WMSDs) are prevalent among masons with 58% reporting lower back pain from repetitive lifting and sustained trunk flexion. Back-support exoskeletons (BSEs) reduce muscle activation and claim to improve posture, yet evidence of their effectiveness in improving posture remains limited. This study investigated whether BSEs improve posture among 23 experienced masons performing masonry tasks, comparing rigid and soft designs. Participants lifted 17.2 kg concrete masonry units with and without exoskeletons. Using YOLO v11 pose estimation to analyze 351,615 video frames, trunk flexion and lateral bend angles were calculated against NIOSH thresholds. Results revealed minimal improvements. Mean trunk flexion reductions of 1.28° and 0.75° were reported, representing only 4.3% and 2.5% reduction in NIOSH risk categories. While BSEs reduce muscle activity, their effectiveness in improving posture is limited. This challenges BSE deployment as intervention for postural correction in masonry work. • Non-intrusive AI-based posture monitoring eliminates need for wearable sensors in construction. • Exoskeletons reduced trunk flexion by only 4.3% and 2.5% of NIOSH risk category increments. • 34.8% of construction workers showed adverse postural responses to exoskeleton use. • Large statistical effects did not translate to meaningful postural improvements. • Study challenges manufacturer claims about universal postural benefits of exoskeletons.