Ergonomic Risk Management Using Machine Learning and Virtual Reality: A Conceptual Framework

Prolonged and repetitive poor postures are often leading to various musculoskeletal dis-orders (MSDs) and significant health risks. Most of the conventional assessment methods rely on periodic evaluations and subjective observations and fall short in providing robust interventions. This paper presents a proposal of an advanced framework that integrates sensor data, machine learning (ML), and virtual reality (VR) for ergonomic risk management. The pilot study observations of the results of this framework using the Random Forest ML algorithm revealed a risk prediction accuracy of $\mathbf{9 8. 5 3 \%}$ for the musculoskeletal risk scores based on predefined posture bounds. Also, whenever an im-proper posture is detected, the system developed in this research can provide immediate feedback through the VR interface that could be useful for real-time corrective guidance so as to prevent or reduce MSDs. Such Artificial Intelligence (AI) and VR enabled interventions are very useful to enhances effective posture corrections for personnel work-ing at construction site operations.