Wearable Network for Multi-Level Physical Fatigue Prediction in Manufacturing Workers - Dataset

This dataset contains data from 43 subjects following two authentic manufacturing protocols and their self-reported fatigue score. The system employs 6 wearable sensors to continuously track vital and locomotive signs from multiple body locations.

The goal of the experiment is to predict fatigue trends in a subject, while they are asked to perform pre-defined manual tasks simulating a manufacturing environment, using data from soft, flexible, wearable sensors and a vision system. The tasks in this study are repetitive and physically exerting involving intricate steps taken in real manufacturing settings. The iterative nature of the tasks facilitates comparative analyses of distinct temporal segments to characterize fatigue. The two manufacturing tasks are (1) Task Composite: Composite Sheet Layup, and (2) Task Harnessing: Wire Harnessing. The task protocol requires the subject to wear sensors to monitor vital and locomotive signs continuously. Additionally, we incorporate a weighted vest to exaggerate the induced fatigue in a reasonable duration for the study to mimic a full shift for a manufacturing worker. Each task consists of two rest periods of 5 minutes each at the start and end, as well as five segments of physical tasks. On average, each task takes a total of 1 hour. Before each data segment, the subject fills out a survey form to indicate their current fatigues as per the Borg scale.