Although maintaining situation awareness (SA) is a critical skill for many complex tasks, there have thus far beenfew rigorous computational approaches to modeling SA behavior and performance. We developed a preliminary computationalmodel of SA focused on remembering the locations of static objects in the visual field. We built this model on the foundationof the ACT-R cognitive architecture, using its declarative memory and vision modules to specify the process of scanning thefield and remembering object locations. In the current work, we demonstrate how this model accounts for human behavior andperformance in two recent experiments: one, a study of object location memory with identities similar to air-traffic control callsigns; and another, a study of remembering the location of shapes of varying size, color, and pattern.

Recent research has shown that when people multitask, boththe subtask structure and the temporal constraints of thecomponent tasks strongly influence people’s task-switchingbehavior. In this paper, we propose an integrated theoreticalaccount and associated computational model that aims toquantify how people balance structural and temporalconstraints in everyday multitasking. We validate the theoryusing data from an empirical study in which drivers performeda visual-search task while navigating a driving environment.Through examination of illustrative protocols from the modeland human drivers as well as the overall fit on the aggregateglance data, we explore the implications of the theory andmodel for time-critical multitasking domains.