An automated knowledge acquisition system for model-based diagnostics

The author discusses how knowledge acquisition will become a key element to improving the design, production, and testing of complex electronic, mechanical, and hydraulic systems. Specifically, knowledge acquisition will become critical to the use of dependency model systems. The author also discusses a means for automating knowledge acquisition and integrating it into the design process, and presents NASA's CAD/CAE knowledge base development tool (KBDT) as a prototype for demonstrating the concept of automated knowledge acquisition. KBDT's basic structure, its operation, and its integration with NASA's knowledge-based autonomous test engineer (KATE), are described. KBDT's unique combination of voice synthesis, voice recognition, natural language processing, and knowledge acquisition processing components, integrated via a blackboard architecture, necessary for this knowledge engineering application, are outlined. >