Research of grinding process fuzzy neural network system based on fuzzy logic

This paper, taking grinding control for example, combines the advantages of fuzzy control and neural network technology, proposes a fuzzy neural network structure based on fuzzy logic rule. Based on fuzzy logic, we give a method in which fuzzy logic can be converted into an equivalent fuzzy neural network. We discuss the network layers structure and derivation process of the algorithm in detail. The simulation results demonstrate the feasibility of the fuzzy neural network. Experiments show that this method are satisfactory, of which the error rate is low, and has a good practical value.

I. Introduction

Fuzzy control and neural network control as two important branches of the field of intelligent control, in many areas have been widely used. But when they are used alone, they have their own drawbacks. However, fuzzy control and neural networks have their own characteristics, we can see that they have significant complementarities. If we can organically combine them, using their respective strengths, avoid their weaknesses, we can get a better system than using fuzzy control system alone or neural network system alone. Fuzzy neural network is such a system that can get fuzzy control and neural network respective advantages of both systems to overcome their own shortcomings.