Development of a sensor-based ergonomic risk assessment framework using machine learning: Application to human-robot collaborative disassembly

doi:10.1016/j.ergon.2026.103943

Development of a sensor-based ergonomic risk assessment framework using machine learning: Application to human-robot collaborative disassembly

2026-04-13

0

PRE

AI

M

Marziyeh Mirzahosseininejad *

DOI：10.1016/j.ergon.2026.103943

Original

Original request for help

Save

Abstract

En 中文

• A sensor-based framework evaluates ergonomic risks in HRC disassembly tasks. • Ten IMU sensors capture real-time motion data for subtask and risk prediction. • RULA and REBA scores label realistic disassembly subtasks across risk levels. • Six ML models are benchmarked; DNN achieves the highest accuracy and F1-scores. • WMSDsNet provides dual-output classification for task type and ergonomic risk.

Keywords:

ergonomic risk assessment

human-robot collaboration

IMU sensors

machine learning

disassembly tasks

Development of a sensor-based ergonomic risk assessment framework using machine learning: Application to human-robot collaborative disassembly

Abstract

Journal

Organization