Application of Model Based Reasoning to Diagnosis of Faults in Inertial Navigation Equipment.

摘要

In 1988 Skinner produced the Blended Diagnostic System (BDS) which was an attempt to provide the Aerospace Guidance and Metrology Center (AGMC) with an expert system capable of diagnosing faults in the Dual Miniature Inertial Navigation System (DMINS). Skinner proposed the blending of a traditional rule-based system with a model-based system. The techniques used to perform the model-based reasoning in BDS are however primitive compared to other techniques currently available. This thesis describes the development of a model-based diagnostic system using techniques pioneered by Randall Davis, and substantially more sophisticated than those used in BDS. A diagnostic prototype for the DMINS was developed which provides a more thorough and consistent diagnosis than does Skinner's model-based system. The models in our prototype were created using the Intelligent Diagnostic Expert Assistant (IDEA) software developed by AI Squared Inc., of North Chelmsford MA. IDEA is based on extensive research by Davis at the Massachusetts Institute of Technology (MIT). Theses. (RRH)