Artificial intelligence techniques, particularly knowledge-based systems, offer the potential for major improvements in diagnostic testing and maintenance. There are applications for knowledge-based systems throughout the design and life cycle of any complex system. Design-for-testability and intelligent built-in test are important aspects of AI applications for future systems or those in early development. There are also many applications for artificial intelligence technology in the diagnosis and maintenance of existing equipment and systems. Expert systems are able to capture and disseminate expertise, integrate a wide variety of knowledge types, including experiential knowledge physical principles, heuristic rules of thumb, and detailed consideration of the symptoms in bringing a wide variety of logical techniques to the diagnostic process. An understanding of structure and function can be used to determine the causes of observed faults. Optical test sequencing and minimization are also greatly facilitated by the use of AI techniques. To date, however, the application of AI to diagnostics has largely been in an experimental off-line fashion. We propose to develop the linking technologies which will allow artificial intelligence systems to access a wide range of standard bus-driven instrumentation and atlas compatible automatic test equipment. These links will provide the basis for the direct application of any number of AI techniques to the solution of real world maintenance problems. In addition, AI offers the potential for fast improvements in the quality and efficiency of programs written in atlas, the most widely used test equipment control language.
