제안 방법

• ADAPT was tested by simulating each pilot's performance on the seven flight maneuvers. A given pilot's knowledge base was accessed by ADAPT, and the model was given desired flight goals and beginning airplane status.
• In this case, a relatively parsimonious model has provided re촤sonable fits to highly complex human performance in a dynamically changing environment. In addition, this work highlights the importance and necessity of turning toward the building of predictive individual models of human performance, accounting for differences at the participant level, rather than simply describing for aggregate performance.
• Piloting performance data derived from student and instructor pilots during simulated flight was used to build and test ADAPT model of pilot instrument-scanning and control performance.
• The data derived from empirical study were used to measure the match between human and model performance. Prior to comparison, we will outline the ADAPT model of piloting skill.
• This information match what was given to individual human pilots at the start of a given flight maneuver. The model executed a Construction-Integration cycle, and selected the most activated plan element whose preconditions were met in either world or general knowledge to fire.
• situation and piloting expertise. This was accomplished by simulating the performance of twenty-five pilots on seven segments of flight, and then comparing human and model performance data to determine ADAPT's predictive validity.