Framework for Approximate Time Rough Control Systems: A Rough-Fuzzy Approach

This paper presents an approach to the design of a framework for an adaptive approximate time rough controller. The class of rough control systems described in this paper utilizes a new form of clock called an approximate time window (atw). Each atw is wired to a single-layer fuzzy neural network constructed with logic-processing neurons. Granulated measurements of the speed and quality of controller behavior provide input to a fuzzy neural network, which is calibratable. Controller rules are derived from a decision table filled with network values. A rough fuzzy Petri net model of such a controller is included. An example application of the this form of adaptive rough controller is given. Copyright© 1998 IFAC.