Development of sleep prediction models for crew fatigue analyses – Crew fatigue modeling


  1. Peng, H.
  2. Bouak, F.
  3. Wang, W.
  4. Lamb, M.
  5. Chow, R.
Corporate Authors
Defence Research and Development Canada, Toronto Research Centre , Toronto ON (CAN)
This report summarizes our work to compare and further develop Sleep Prediction Models (SPMs) for crewing analysis with the integration of a DRDC Fatigue Model (DFM) into a DRDC naval crewing analysis tool called Simulation for Crew Optimization and Risk Evaluation (SCORE). We identified three sleep prediction models and then programmed them using MATLAB®, a generic computer programming tool. We compared these models and the AutoSleep module in Fatigue Avoidance Scheduling Tool (FAST), commercial software, with empirical data from the literature and our own at-sea trial. We also added the SPMs to our DFM and compared crew cognitive performance predicted from both self-reported and model-predicted sleep. We found better predictions by SPMs than AutoSleep for both sleep and performance as indicated by their mean square errors against the literature data and our self-reported sleep data. We further improved sleep predictions for a Navy operational setting involving various work/rest schedules such as watchstanding schedules by optimizing some of the model parameters. The flexibility of our model would allow adjustment of parameter values and additions of new algorithms and mechanisms for other factors such as circadian phase shifts, and thus improve model predictions and usefulness for both civilian and military applications. Future development and validation of our SPM and DFM for crew fatigue analyses were warranted.

Il y a un résumé en français ici.

Sleep model;work schedule;crewing analysis
Report Number
DRDC-RDDC-2017-R127 — Scientific Report
Date of publication
01 Nov 2017
Number of Pages
Electronic Document(PDF)

Permanent link

Document 1 of 1

Date modified: