A preliminary study of hybrid fuzzy/statistically-based controllers

Abstract

This paper describes the development and preliminary analysis of hybrid fuzzy/statistically-based controllers, including Fuzzy CUSUM. The Fuzzy CUSUM controller is a modification of a CUSUM type controller developed from statistical quality control concepts with fuzzy elements added. These controllers may be used in several process applications, especially the quality control of batch operations. First, the original controller is described and its performance evaluated. Next, the fuzzy elements are added to the controller and its performance is detailed. A completely randomized block experimental design and Duncan's multiple range test are used to compare the hybrid controllers to other types of controllers currently being explored, under a variety of disturbance conditions. The hybrid fuzzy/statistically-based controllers significantly outperform all other controllers examined.<>