Self-improving ATE

The University of Louisville, under contract from the Naval Ordnance Station/Louisville (NOSL), was given the task of exploring the use of artificial intelligence to increase the efficiency of automated test equipment (ATE) used to diagnose faults in low-frequency analog/digital circuit card assemblies (CCA). The authors present an account of the realization of level 1 of a multilevel project organized to accomplish this task. Level 1 concentrates on the design and development of an expert system to guide the ATE through fault-isolation testing. Intelligent computer-aided testing (I-CAT), an expert system, was selected as the nucleus for prototype development and deployment. The primary goal of level 1 was to reduce ATE fault-isolation run time. The static decision, tree previously used to control the ATE was replaced by the dynamic diagnostic strategy of I-CAT, which updates its reasoning by taking into account previous failure rates and reliability data. Node measurements within ambiguity groups are applied to a pattern recognizer to help identify component failures within individual ambiguity groups