Controller Performance Assessment (CPA) of Intelligent Control for Non-linear System

Abstract

The objective of this paper is to design and undertake comparative analysis of classical and intelligent
controllers for nonlinear system. These controllers are compared based on controller performance
assessment in which the different parameters as overshoot, steady-state error, rise time, settling time,
response of reference change and output variance are analyzed. To achieve these objectives, the water tank
control problem as nonlinear system has been built in Simulink and implementations traditionally classical
controller and advance intelligent controller as fuzzy logic and neural network are performed. In classical
controller, the proposal-integral control law is used as control algorithm. Sugeno model is used for FL
controller, while back propagation algorithm is used to train the NARMA-l2 neural network control. Three
controllers are designed and their performances are compared.
Keywords: Controller performance assessment (CPA), classical controller, fuzzy logic controller, NARMAl2
neural controller, nonlinear system