INFLITE: An Intelligent Instrument Flight Trainer with Computer-Generated Speech

Researchers at the Air Force Human Resources Laboratory AFHRL are applying a taxonomy of learning skills Kyllonen Shute, 1987 to the pedagogical issues surrounding Intelligent Tutoring System ITS design and development. The taxonomy provides us with a means of categorizing target domains and consequently specifying the appropriate training approaches for particular ITSs. Further, it highlights classes of domains for which ITSs are appropriate but have not yet been developed. This paper focuses on an application in an area that has only recently been explored a class of tasks which we refer to as high-performance tasks Regian Shute, 1988. In high- performance tasks, there is greater requirement for highly speeded, reliable, and automatic task performance than is found in the typical knowledge-rich ITS domains e.g., medical diagnosis, electronic troubleshooting. This paper describes an Instrument Flight Trainer INFLITE developed at AFHRLs Training Systems Division to test the concept of using artificial intelligence to train high-performance tasks. INFLITE is in no sense a serious attempt to develop an application-ready instrument flight training device, Rather, INFLITE is an experimental system designed for the purpose of evaluating a promising approach to training high-performance tasks. Our decision to use flight simulation as the prototype domain was guided by a desire to use an inherently interesting task to increase motivation in our experimental subjects.