UMBC ebiquity
First IEEE Conference on Artificial Intelligent Applications
FOREST - An Expert System for Automatic Test Equipment
December 1, 1984
This paper describes FOREST, an expert system for fault isolation and diagnosis in the Automatic Test Equipment (ATE) domain. Current ATE systems can correctly handle 90 to 95 percent of faults, but the residue accounts for a considerable cost in terms of equipment downtime and the human expert's time. FOREST is an attempt to handle the residue of hard problems with current expert systems techniques. In particular, the incorporation of an AI approach allows us to handle two serious problems that the existing decision tree techniques can not: problems involving multiple faults and problems caused by components or systems that gradually drift out of calibration. FOREST is implemented in PROLOG and has an architecture that combines an object-oriented representation language (FIR), a general inferencing mechanism (PROLOG), an inferencing engine for reasoning with certainty factors (PINE), and an explanation generation system (ELM).
InProceedings
IEEE
Downloads: 155 downloads