System-level applications of model-based diagnostic reasoning

Abstract

Progress in developing adequate diagnostic capabilities for future computing and communications systems has been hampered by the lack of an adequate technological basis. Over the past decade, researchers within the artificial intelligence community have developed model-based diagnostic reasoning techniques. These techniques have the potential to provide comprehensive diagnostic coverage and to automate the design analysis process. MITRE is developing a prototype of the Generic Model-based Diagnostic System (GMODS) that is able to use VHDL models of circuit structure and behavior to diagnose single and multiple points of failure within digital sequential circuits. We provide an overview of GMODS and discuss the remaining research issues that need to be addressed before model-based reasoning systems such as GMODS will be of general utility to the R&M community. We also explore three potential system-level application areas-design analysis, depot-level diagnostics, and embedded diagnostics. We conclude that design analysis and depot-level diagnostics represent the most promising applications during the next few years.<>