Effects of fatigue on simulated space telerobotics performance : a preliminary study analysis

Abstract

Astronauts on the International Space Station must perform mission critical space telerobotics tasks consistently despite restricted and slam shifted sleep and circadian schedules and long session durations, which all potentially degrade cognitive function, response time, and attention. A ground laboratory experiment was designed to 1) Determine the effects of fatigue on performance metrics in simulated space telerobotics tasks, 2) Examine the relationships between performance on complex robotics tasks and traditional metrics of cognitive task performance and sleepiness, 3) Assess the efficacy of caffeine and blue light countermeasures, and 4) Evaluate individual subject vulnerability to fatigue. Subjects were screened for robotics aptitude, and trained on three robotics tasks and a mental workload assessing secondary task at MIT. After a week of 6-hr sleep restriction, they were admitted to the Brigham and Women's Hospital sleep laboratory, a time-cue free environment, and underwent a 13 day double blind protocol including physiologic monitoring and robotic and cognitive testing. Their sleep schedule was repeatedly slam shifted 9 hours earlier then they performed the robotics tasks under different countermeasure conditions. This thesis documents the protocol and details of the robotics training and testing, and includes a preliminary analysis of data from 16 subjects focusing on robotics and secondary task data. Since the study is continuing and investigators are blinded to countermeasure conditions, data from the countermeasure sessions is not included. Thesis goals were to 1) Analyze the predictive capability of spatial ability tests on individual robotics performance, 2) Evaluate the effects on robotics metrics of proxy measures of circadian and time-on-task, and 3) Assess individual differences in performance and vulnerability to fatigue. The Vandenberg Mental Rotation Test was found to be the best predictor of both robotics screening test and experimental performance, although an average of four spatial ability tests was slightly better for screening purposes. A comparison between a final training and non- countermeasure test session indicated that slam shifting had no significant effect on group average performance in any of the three robotics tasks or the secondary task. However, within the slam shifted session, a time-on-task related effect in secondary task performance was evident, suggesting that mental workload gradually increased even though subjects were able to maintain primary robotics task performance. Inter-subject differences were consistently larger than other effects.