A hybrid knowledge representation for the domain model of intelligent flight trainer

Abstract

After nearly 40 years of development, Intelligent Tutoring Systems (ITS) is moving from research labs into the real world. However, Flight training seems to be somewhat divorced from the mainstream of ITS' application. One reason is the difficulty in the knowledge representation of flight training domain, especially the procedural knowledge of flight maneuvers. After a brief review of knowledge representation in ITS, it is clear that the requirements of KR are naturalness and ease to update. Concept Knowledge Tree (CKT) is an ideal method to represent structural and relational knowledge. Behavior Tree is an elegant and understandable solution to procedural knowledge representation for creating reusable logic. CKT and BT can be combined through script and thus a powerful hybrid knowledge representation method was put forward. With authoring tools, a domain expert can even work without the help of knowledge engineer. The analysis of normal takeoff as an example of a single flight maneuver and the construction of flight maneuver knowledge tree proved the hybrid method an appropriate method to develop the domain model of intelligent flight trainer.