Self-Tuning Neural Controller with Orthonormal Activation Function

Abstract

This paper presents a novel methodology for modeling and control of nonlinear processes using artificial neural network with orthonormal activation functions (e.g. harmonic, Legendre, Laguerre and Chebychev). An algorithm for modeling and self-tuning control of uncertain nonlinear complex processes is proposed. Several types of orthonormal functions for nonlinear process modeling and control are analyzed and compared. Simulation results are provided to illustrate high perfonnance of artificial neural networks with the orthonormal activation functions in practical industrial problems. The obtained results are reviewed and potential shortcomings of the analyzed methods are discussed.