Intelligent Simulation-Based Tutor for Flight Training
Abstract
Today's military flight simulators have dramatically reduced the cost of training by providing cheaper, effective alternatives to training on a real aircraft. However, flight training is still limited by the availability of instructor pilots. The adage "practice makes perfect" is nowhere truer than in the learning of psychomotor skills such as flying. Ideally, trainees should be able to practice flying skills on their own to complement instructor-led training. However, most flight simulators do not have any automated assessment and tutoring facilities, making them ineffective as self-paced learning environments. The Army has funded pioneering research on developing automated tutors for flight training, specifically for training initial-entry rotor-wing pilots. An early rule-based system, called the Intelligent Flight Trainer (IFT), monitored trainees' flight performance and provided adaptive coaching. It provided instructional assistance by regulating the challenge level of a flight task, and through overt spoken feedback to inform trainees when they were flying out of range of specified flight parameters. Evaluations showed that while this system was effective in improving flying skills, it was inflexible in terms of it assessment and instruction strategies. The Army is currently funding research on a next-generation automatic flight trainer, called AIS-IFT, that improves upon the IFT. AIS-IFT is designed to be flexible and extensible in terms of assessment and tutoring procedures. A visual authoring tool lets Subject Matter Experts (SMEs) and course designers modify or create powerful instructional behavior with little programming effort. Whereas the previous effort had the instructional approach embedded deep in the tutoring system, the new approach separates the specific instructional strategies from the ITS infrastructure, thus empowering SMEs and course authors to create a tutor with pedagogy that is customized to their domain.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 2004
- Accession Number
- ADA439800
Entities
People
- Daniel Fu
- Emilio Remolina
- Richard Stottler
- Sowmya Ramachandran
- William R. Howse
Organizations
- Stottler Henke Associates