Hybrid diagnostic strategy for an expert system controlled automatic test system (EXATS)

A description of EXATS is given. A knowledge-based system is interfaced to an automatic test system (ATS) to provide a flexible test sequence, controlled by the actual fault signature. Successful diagnoses are fed back to enhance the knowledge base (KB). The reasoning process for test selection and test evaluation consults two separate KBs. The experiential knowledge about the heuristic symptom fault relation is organized as a hierarchy of rule sets which are exploited by a top-down refine strategy. The physical knowledge about structure and function of the unit under test is modeled as a logical function model. Continuous signal variables are treated as discrete events, allowing the representation of the transfer functions as a state variable truth table. Thus, an implicit fault model is obtained. A concept using Prolog to implement the propagation of the state variables and their corresponding time intervals is introduced. The diagnosis is based on ambiguity sets of suspect components for each failed test. These sets are generated by successive constraint deactivation. EXATS was demonstrated with a line replaceable unit on the radio frequency automatic test equipment.<>