Intelligent Tutoring for Diagnostic Problem Solving in Complex Dynamic Systems
Abstract
Maintenance training for diagnostic problem solving in complex dynamic systems is carried out either on the job or on simulators. When simulators are used for training, their effectiveness can be improved by integrating intelligent tutoring systems (ITS) into the training programs. Research results from ITSs developed for simpler task domains are generally not very useful in complex engineered domains due to lack of appropriate knowledge representation techniques. The focus of our research is the development of a methodology for decomposing, organizing, and representing domain knowledge of complex dynamic systems for building functional computer-based intelligent tutors. Using our knowledge representation methodology, we implemented an ITS on an Apple Macintosh II computer for the marine power plant domain. The ITS is comprised of a simulated power plant, the tutor, and mouse-based direct manipulation graphical interfaces. The ITS was experimentally evaluated using Naval ROTC cadets as subjects. Performance of the subjects was analyzed using measures such as percentage of premature and correct diagnosis and percentage of relevant and irrelevant diagnostic tests were used. Results show that a simulator alone is inadequate, whereas a simulator in conjunction with an ITS can help develop efficient troubleshooting skills.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1991
- Accession Number
- ADA242619
Entities
People
- Vijay Vasandani
Organizations
- Georgia Tech