Study on the structure and algorithm of a feedback fuzzy controller

This paper presents the structure and algorithm of a new fuzzy logic controller which contains a feedback of control action in its input and therefore is called dynamic fuzzy controller, mainly studies the first-order dynamic fuzzy controller and six new corresponding fuzzy inference rules are given. The results show that the performances of the control systems with the first-order dynamic fuzzy logic controller, especially the robustness of resisting uncertainty and anti-time-delay, are much better than those with the traditional static fuzzy controller which has two inputs and forty nine fuzzy inference rules.

I. INTRODUCTION

The theory of fuzzy set has been in developing for 30 years since it was presented by Zedeh in 1965, and one of its most important application fields is in control engineering^{[1], [2], [3], [4]}. Generally speaking, fuzzy logic controller will be an effective way in designing a control system when controlled processes are too difficult to be described by physical or chemical principles, or its models are too complicated to be solved. The design of a fuzzy controller does not depend on the analysis and synthesis of process mathematical model, but all of its control rules come from the expert's knowledge or the operator's experiences^{[3], [4], [5]}, and therefore fuzzy logic controller is not dynamic in generally^[4]. We can also know that fuzzy logic controller is a static controller from that fuzzy controller is equivalent to a multi-value relay in stability analysis [3].