A Framework for the Design of a Voice-activated, Intelligent and Hypermedia-based Aircraft Maintenance Manual
Date of Award
Doctor of Philosophy in Computing Technology in Education (DCTE)
Graduate School of Computer and Information Sciences
Getrude W. Abramson
Marlyn Kemper Littman
Michael J. Laszlo
Federal Aviation Regulations require Aviation Maintenance Technicians (AMTs) to refer to approved maintenance manuals when performing maintenance on airworthy aircraft. Because these manuals are paper-based, larger the size of the aircraft, more cumbersome are the manuals. Federal Aviation Administration (FAA) recognized the difficulties associated with the use of large manuals and conducted studies on the use of electronic media as an alternative to the traditional paper format. However, these techniques do not employ any artificial intelligence technologies and the user interface is limited to either a keyboard or a stylus pen. The primary emphasis of this research was to design a generic framework that would allow future development of voice-activated, intelligent, and hypermedia-based aircraft maintenance manuals. A prototype (VIHAMS-Voice- activated, Intelligent, and Hypermedia-based Aircraft Maintenance System) was developed, as a secondary emphasis, using the design and development techniques that evolved from this research.
An evolutionary software design approach was used to design the proposed framework and the structured rapid prototyping technique was used to produce the VIHAMS prototype. Voice Assist by Creative Labs was used to provide the voice interface so that the users (AMTs) could keep their hands free to work on the aircraft while maintaining complete control over the computer through discrete voice commands. Knowledge Pro for Windows 1M, an expert system shell, provided "intelligence" to the prototype. As a result of this intelligence, the system provided expert guidance to the user.
The core information contained in conventional manuals was available in a hypermedia format. The prototype's operating hardware included a notebook computer with a fully functional audio system. An external microphone and the built-in speaker served as the input and output devices (along with the color monitor), respectively. Federal Aviation Administration estimates the United States air carriers to operate 3,991 large jet aircraft in the year 1996 (FAA Aviation Forecasts, 1987-1998). With an estimate of seventy manuals per such aircraft, the development of intelligent manuals is expected to impact 279,370 manuals in this country. Soon, over 55 thousand maintenance technicians will be able to carry the seven pound system to an aircraft, use voice commands to access the aircraft's files on the system, seek assistance from the expert system to diagnose the fault, and obtain instructions on how to rectify the fault.
The evolutionary design approach and the rapid proto typing techniques were very well suited for the spiral testing strategy. Therefore, this strategy was used to test the structural and functional validity of this research. Professors Darrell Anderson and Brian Stout (Aviation faculty at San Jose State University) and Mr. Gregory Shea (a United Airlines mechanic and SJSU student) are representatives of the real-world users of the final product. Therefore, they conducted the alpha test of this prototype. Mr. Daniel Neal and Mr. Stephen Harms have been actively involved in light aircraft maintenance for more than ten years. They evaluated the prototype's usability. All the above evaluators used standard testing tools and evaluated the prototype under field conditions.
The evaluators concluded that the VIHAMS prototype used a valid fault diagnosis strategy, the system architecture could be used to develop similar systems using off-the shelf tools, and the voice input system could be refined to improve its usability.
Manoj S. Patankar. 1997. A Framework for the Design of a Voice-activated, Intelligent and Hypermedia-based Aircraft Maintenance Manual. Doctoral dissertation. Nova Southeastern University. Retrieved from NSUWorks, Graduate School of Computer and Information Sciences. (766)