Musculoskeletal Risk Modeling Using Sensors and Machine Learning

This study develops a predictive modeling approach for occupational ergonomic risk in manual lifting tasks to enhance real-time biomechanical risk assessment, aimed at mitigating occupational musculoskeletal disorders (MSDs). Wearable Electromyography (EMG) sensors were used to collect data on muscle activity of eight upper extremity muscles across ten participants (n = 10) engaged in both low-risk and high-risk tasks for MSDs. A machine learning model was developed and optimized for lifting task risk classification. The model shows high accuracy, precision, and recall values in classifying low and high-risk lifting tasks. This approach significantly surpasses traditional occupational risk assessment methods, which often rely on historical data and are reactive rather than proactive. The integration of wearable sensors with machine learning provides precise risk classification and facilitates the implementation of effective safety interventions across various occupational settings. This strategy has potential to improves safety planning and can contribute to substantial reduction in the occurrence and severity of MSDs, with ultimate goal of enhancing overall workplace safety and health.