Simulation-Based Evaluation of Adaptive Automation Revoking Strategies on Cognitive Workload and Situation Awareness

Adaptive systems have used a variety of approaches to return the task responsibility back to the user, including restoration of performance or physiological thresholds, leveling of task loads, and user-initiated deactivation. Little attention has been paid to evaluating the relative effectiveness of these various handoff mechanisms, or the impact that these mechanisms may have on operator workload and situation awareness (SA). This study uses an Improved Performance Research Integration Tool simulation, based upon a human-in-the-loop study, to examine two different revoking strategies (minimum duration and workload threshold) across numerous levels in order to identify impacts on operator workload and SA. This study finds that operator workload and SA are affected by the type of revoking strategy as well as the particular level used for that strategy. For the scenario examined, automation revoking that relied upon a minimum duration of at least 5 s was found to be more effective in reducing workload and increasing SA than using a revoking strategy based upon workload thresholds.