Using non-invasive wearable sensors to estimate perceived fatigue level in manual material handling tasks

Abstract

Physical fatigue in manual material handling (MMH) may cause musculoskeletal disorders (MSDs), which threatens the well-being of workers. However, conventional techniques for measuring fatigue have their limitations and are difficult to implement in realistic working conditions without sufficient resources. In this study, we proposed a method utilizing non-invasive wearable sensors to collect bio-signals (respiration, photo-plethysmography, and electrodermal activity) and estimate perceived physical fatigue. Six participants volunteered in two MMH tasks at two paces. Subsets of five bio-signal measures were selected to estimate perceived fatigue levels using a universal regression model and six individualized regression models. We classified perceived fatigue into three levels and examined the correct classification rate of the estimated fatigue levels. Correct classification rates for the general model and the individualized models were 67% and 80%, respectively. Results confirm the feasibility to predict fatigue level using wearable sensors, but the regression models should be used with caution.