TABLE OF CONTENTS The aim of this chapter is to introduce background concepts in nonlinear systems identification and control with artificial neural networks. As this chapter is just an overview, with a limited page space, only the basic ideas will be explained here. The reader is encouraged, for a more detailed explanation of a specific topic of interest, to consult the references given throughout the text. Additionally, as general books in the field of neural networks, the books by Haykin [1] and Principe et al. [2] are suggested. Regarding nonlinear systems identification, covering both classical and neural and neuro-fuzzy methodologies, [3] is recommended. [4] and [5] should be used in the context of B-spline networks.
The main goal of systems identification is to determine models from experimental data. These models can afterwards be employed for different objectives, such as prediction, simulation, optimisation, analysis, control, fault detection, etc. Neural networks, in the context of system identification, are black-box models, meaning that both the model parameters and the model structure are determined from data.
Model construction is an iterative procedure, often done in an ad hoc fashion. However, a sequence of steps should be followed, in order to decrease the number of iterations needed to obtain, in the end, a satisfactory model [3]. As a first step, with the help of any prior knowledge available, the relevant inputs for the model must be identified. Afterwards, as these...
