Cognitive Probe Project: Development of a Testbed for Collecting Cognitive Model Parameterization and Validation Data

摘要

The Air Force Research Laboratory's Human Performance Model Integration (HPMI) program is exploring the merit and feasibility of combining multiple human performance models possessing dissimilar architectures to create integrated representations of human behavior that address application-specific requirements for model fidelity while controlling cost. The first exploration of HPMI feasibility involves the integration of a task-network model of a strike fighter pilot with an Adaptive Control of Thought-Rational (ACT-R) cognitive model. The ACT-R cognitive model is to replace the strike fighter pilot model's implementation of shootlist management, which is a dynamic and cognitively- intensive target-prioritization task. The Cognitive Probe project described in this report developed a concept and a virtual simulation testbed for obtaining the context-specific parameterization data needed to populate the ACT-R shootlist management model. One observation based on the data collection was that the earlier validation of the initial implementation of the strike fighter pilot model, which compared mission outcomes and task performance of the model against that of humans, was not sufficiently sensitive to cognitive aspects to pick up differences in shootlist management search strategies. Cognitive testbeds such as this may be beneficial for defining or verifying effective tactics and interface configuration before developing human performance models.