Master of Science in Industrial Engineering (MSIE)


Mechanical Engineering

Document Type



When a person continues to work despite being fatigued, an accident may occur or productivity may decrease. Reducing fatigue related accidents is complicated due to the subjective nature and pervasiveness of the feeling of fatigue. Because finding one universal definition or measure of fatigue is problematic, fatigue is typically measured by a combination of factors such as amount of sleep obtained, physical or cognitive work performed, or sleepiness which can then be related back to a measure of performance. Until recently, sleep measurement was restricted to a lab setting; however, with the emergence of actigraph devices, it is now possible to measure sleep in a person’s natural environment. Therefore, the purpose of this research is to determine if operator performance on a given day can be predicted based on the amount of sleep obtained the previous night as measured on an actigraph. To do this, the researchers used the Consensus Sleep Diary (CSD) and the Basis Band (BB) actigraph to subjectively and objectively measure the Total Sleep Time (TST), Sleep Efficiency (SE), and Number of Awakenings (NOW) recorded over three nights. Then, performance was measured using the ten-minute Psychomotor Vigilance Self-Test (PVT) within one to three hours of waking up and prior to performing any significant activities (Basner & Dinges, 2011). The two measures from the PVT are 1/RT (RT=mean reaction time) and Number of Lapses (NOL) (Basner & Dinges, 2011) which were correlated against the three sleep measures. Additionally, the three sleep measures were correlated to a subjective rating of a person’s feeling of fatigue using the Stanford Sleepiness Scale (SSS). The Total Sleep Time and Number of Awakenings as measured on the Basis Band were significantly different from the same measures reported using the Consensus Sleep Diary. When the sleep measures were correlated against the performance measures, the only two statistically significant results were the correlation between the Total Sleep Time (on both the BB and CSD) to the Number of Lapses (ñ=-.1919, p= .0446 and ñ=-.2168, p= .0229, respectively) and the Total Sleep Time (on both the BB and CSD) to the Stanford Sleepiness Scale rating (ñ=-.2059, p= .0309 and ñ=-.2702, p= .0043, respectively). These results indicate that the Basis Band and the PVT are not sensitive enough to be used as predictors of operator performance.



Document Availability at the Time of Submission

Release the entire work immediately for access worldwide.

Committee Chair

Harvey, Craig