Evenness in Two-End Loads of Padders by Genetic-Based Self-Tuning Fuzzy Control

Abstract

This paper presents an experimental study of self-tuning fuzzy control of the two-end loads of padders. A simplified padder set-up is used. In conventional fuzzy control, membership functions and control rules are designed based on human subjective criteria, so very often, the design attempt may not lead to excellent performance. Our membership functions and control rules for load control of padders are experimentally investigated in this application. However, the ability to maintain two-end loads at the desired values is not satisfactory. To improve performance, membership functions are tuned based on an optimization technique of genetic algorithms to yield self-tuning fuzzy control. The experimental results indicate that the self-tuning fuzzy controller can maintain the two-end loads not only evenly and at the desired values. The tuning scheme is helpful in eliminating time-consuming trial-and-error procedures when refining membership func tions or control rules.