Design of Fuzzy Neural Network Based Multi-Variables Controller for Manipulators

Yoshishige Sato

Abstract


This paper proposes a robust multivariable control design by intelligent control that uses a fuzzy neural network by producing robustness capable of automatically controlling gain against a conventional, fixed PID control system. This structural feature of the proposed controller forms a nonlinear deviation compensator using fuzzy neural networks. Therefore, in multidimensionality the inverse dynamic model portion of the control law is referred to as a linearizing and decoupling control law. This method uses a control law where parameter response leads to critical damping and adaptive changes in gain according to time, making it possible to decouple mutual interference in each multivariable system.

Keywords


fuzzy neural network, multivariable control, decoupling, manipulator control

Full Text:

PDF