Using Empirical Research and Computational Modeling to Predict Operator Response to Unexpected Events

Abstract

This paper describes an effort to model and predict astronaut performance during sudden workload transitions in long duration missions. Our approach to the work is heavily based on empirical research. We have performed a set of meta-analyses 1) to identify the quantitative effects of poor sleep on task accuracy and task completion time, and 2) to develop a model of operator task selection during multitasking. We are currently developing a model, based on a literature review, to predict the effects of automation design factors on operator task performance. This paper gives an overview of the project, presents the overall model of operator performance during a workload transition, and describes the empirical and theoretical underpinnings of a model that predicts the effects of automation design on operator performance.