Cognitive Fatigue Assessment Using Physiological Sensors during Human-Robot Interaction for Activities of Daily Living

A multimodal dataset is presented for the cognitive fatigue assessment of physiological minimally invasive sensory data of Electrocardiography (ECG) and Electrodermal Activity (EDA) and self-reporting scores of cognitive fatigue during HRI. Data were collected from 16 non-STEM participants, up to three visits each, during which the subjects interacted with a robot to prepare a meal and get ready for work. For some of the visits, a well-established cognitive test was used to induce cognitive fatigue. The developed cognitive fatigue assessment framework filtered noise from the raw signals, extracted relevant features, and applied machine learning regression algorithms, such as Support Vector Regression (SVR), Gradient Boosting Machine (GBM), and Random Forest Regressor (RFR) for estimating the Cognitive Fatigue (CF) level.