Detection of task type through unobtrusive physiological monitoring

Date

7/10/2022

Journal Title

Journal ISSN

Volume Title

Publisher

51st International Conference on Environmental Systems

Abstract

As deep space missions become more autonomous and self-reliant, the need to execute tasks collaboratively between humans and robots increases. As such, it also becomes important to monitor crew workloads in real time in order to adjust task allocation or advise countermeasures. In this work, we demonstrate the utility of monitoring psychophysiological signals � electrocardiogram (ECG), electrodermal activity (EDA), etc. � as an unobtrusive method for detecting the type of task being performed and level of effort expended by the subject. We conducted testing with 13 participants over periods of 8-10 days. The participants completed the same daily protocol with two different kinds of rest activities and two different task types. The initial rest phase consisted of viewing meditative videos while slowly riding a stationary bike, with the next two phases viewing the videos only; the tasks consisted of a computer-based math test coupled with either slow biking or fast biking. Assessments of performance were measured during each task, and survey measures were collected between tasks. These observational and subjective measures are analyzed to identify and characterize trends shown in the psychophysiological signals. Results show that heart rate variability (HRV) decreases and mean EDA levels increase with increasing cognitive load, which coincides with other published findings. Additionally, individual participant analyses revealed that strong cognitive exertion prior to the experimental protocol resulted in no change in physiological signals throughout the testing, suggesting that this may be an indication of cognitive over-exertion. Such an indicator could be used for corrective action during real-time operations and provide an important capability in the development of collaborative human-robot teams.

Description

Katya Arquilla, Massachusetts Institute of Technology, US
Michael Zero, University of Colorado Boulder, US
Kaitlyn Hauber, University of Colorado Boulder, US
Mark Shelhamer, Johns Hopkins University, US
David Klaus, University of Colorado Boulder, US
Christine Fanchiang, The Space Research Company, LLC, US
ICES405: Human/Robotics System Integration
The 51st International Conference on Environmental Systems was held in Saint Paul, Minnesota, US, on 10 July 2022 through 14 July 2022.

Keywords

cognitive performance, psychophysiology, biosignals, human-robot interaction

Citation