Fuzzy integrated process supervision

The integrated process supervision (IFS) is a structured refinement of the expert control architecture that allows the integration of different control techniques within a single generic framework. This is achieved in IFS through the classification of system behaviour into acceptable, malfunction and faulty behaviours according to certain system performance measurement using a simple thresholding, detection mechanism. This technique raises two important research issues, namely, the choice of the thresholds between the classes of system behaviours, and the sensitivity of such thresholding mechanism to uncertainty and noise during the determination, of system behaviours. This paper attempts to address the second issue by using fuzzy sets instead of crisp behaviour categorisation to represents system behaviours. Extensive experimental results have shown that the use of fuzzy sets provides significant noise tolerance in the switching mechanism and improves the transparency of the integrated process supervisor, as well as the modularity of the IFS scheme