Applying Engineering Models of Human Behavior to Space Flight Performance

Abstract

Most complex human-device interactions that arise in realistic applied settings are heterogeneous along several dimensions that significantly restrict the scope and effectiveness of traditional modeling tools. In particular, different subcomponents of the same overall task may vary significantly with respect to practice level and speed-accuracy tradeoffs. The goal of this paper is to lay out the path toward a set of technologies and underlying psychological theory that will permit rapid modeling of such heterogeneous tasks, and therefore rapid evaluation of proposed interface and task structure designs. It will be critical to make progress toward the goal in three specific ways: (1) Development of simple, high-level languages that permit the rapid specification of new tasks composed hierarchically from existing task and cognitive architectural building blocks; (2) Development of the technology and underlying psychological theory that will permit the system engineers to specify the degree of learning or skill associated with separate subcomponents of the overall task, and derive the behavioral consequences of those skill assumptions without simulating the learning process itself; and (3) Develop the technology and theory that will permit the system designers and engineers to specify speed-accuracy tradeoffs separately for subcomponents of the overall task, and derive the behavioral consequences of those assumptions in concert with the skill assumptions. The outcome of such efforts will support the development of modeling tools and design guidelines for mission systems based on common skill sets and allow prediction of learning and performance time for such systems. Directly supporting two central Human-Systems Integration themes, this paper will focus on modeling tools and models needed to support prediction of degrading human performance over the duration of reference missions toward informing the design of systems to mitigate these effects.