Document Type


Publication Date





This paper describes an experiment that was conducted to provide an empirical foundation for estimation of parameters for air traffic controller performance modeling efforts presently pursued within the NASA DAG-TM CE-6 model development. The focus of the work was the task prioritization scheme used in these models. A total of 11 retired FAA controllers and supervisors assigned to the FAA Technical Center volunteered to participate in the experiment. A part-task experimental simulation that presented the participating controllers with several simultaneous tasks in four quadrants, or panes, on a single display was used. Only one pane and typically one task could be viewed at a time. This allowed for measurement of controllers’ attention to each task. All events unfolding in the experimental scenarios and controllers’ actions were recorded and timed as well. From these data, several dependent variables were derived, focusing on the temporal aspects of controllers’ performance and their prioritization of simultaneously available tasks. The results indicate that taskload was manipulated successfully and resulted in measurable differences between experimental conditions in both taskload and performance, the latter evinced by the time elapsed in a window of opportunity for a given task before action was taken on it as well as time remaining in the window of opportunity when action was completed. However, it appears that either the controllers were not aware of these temporal features of their tasks or that other factors dominated their prioritization decisions. Task prioritization may hence be driven by task characteristics that are categorical rather than continuous and quantifiable.