Taguchi-tuned radial basis function with application to high precision motion control

Abstract

This paper presents a novel application of Taguchi method to systematically tune the weights of a radial basis function (RBF) network, which is widely used for modelling vaguely defined but smooth nonlinear functions. The main strength of this method is the well-defined and systematic statistical design procedure, which is amenable to practical implementation. To illustrate the effectiveness of the Taguchi-tuned RBF, a test platform is required. This approach is applied to a platform involving high precision motion control. The developed method then is used to tune a composite motion controller incorporating RBF-based adaptive control in a high precision motion environment. Simulation and experimental results reveal the effectiveness of a Taguchi-tuned RBF.