An improved technique for applying fuzzy logic in a model-based diagnostics reasoner

The authors apply an improved technique based on fuzzy logic to fulfill the requirement to compute a fault hypothesis in a model-based diagnostic reasoner. The primary focus of the paper is to extend previous work and illustrate the use of the improved algorithm. The paper deals primarily with two types of uncertainty: test outcome uncertainty and inference uncertainty. The issue of conflict detection and conclusion rehabilitation is addressed. This proposed fuzzy logic approach uses the confidence parameters defined in the IEEE 1232: Artificial Intelligence Exchange and Service Tie to All Test Environments (AI-ESTATE) standard to provide a robust reasoning under uncertainty capability.