A Closed-Loop Model of Operator Visual Attention, Situation Awareness, and Performance Across Automation Mode Transitions

Objective:

This article describes a closed-loop, integrated human–vehicle model designed to help understand the underlying cognitive processes that influenced changes in subject visual attention, mental workload, and situation awareness across control mode transitions in a simulated human-in-the-loop lunar landing experiment.

Background:

Control mode transitions from autopilot to manual flight may cause total attentional demands to exceed operator capacity. Attentional resources must be reallocated and reprioritized, which can increase the average uncertainty in the operator’s estimates of low-priority system states. We define this increase in uncertainty as a reduction in situation awareness.

Method:

We present a model built upon the optimal control model for state estimation, the crossover model for manual control, and the SEEV (salience, effort, expectancy, value) model for visual attention. We modify the SEEV attention executive to direct visual attention based, in part, on the uncertainty in the operator’s estimates of system states.

Results:

The model was validated using the simulated lunar landing experimental data, demonstrating an average difference in the percentage of attention ≤3.6% for all simulator instruments. The model’s predictions of mental workload and situation awareness, measured by task performance and system state uncertainty, also mimicked the experimental data.

Conclusion:

Our model supports the hypothesis that visual attention is influenced by the uncertainty in system state estimates.

Application:

Conceptualizing situation awareness around the metric of system state uncertainty is a valuable way for system designers to understand and predict how reallocations in the operator’s visual attention during control mode transitions can produce reallocations in situation awareness of certain states.

http://ift.tt/2c0UBYQ