Nonlinear System Identification

Abstract

PrefaceThe goal of this book is to provide engineers and scientIsts in academia and industry with a thorough understanding of the underlying principles of nonlinear system identification. The reader will be able to apply the discussed models and methods to real problems with the necessary confidence and the awareness of potential difficulties that may arise in practice. This book is self-contained in the sense that it requires merely basic knowledge of matrix algebra, signals and systems, and statistics. Therefore, it also serves as an introduction to linear system identification and gives a practical overview on the major optimization methods used in engineering. The emphasis of this book is on an intuitive understanding of the subject and the practical application of the discussed techniques. It is not written in a theorem/proof style; rather the mathematics is kept to a minimum and the pursued ideas are illustrated by numerous figures, examples, and real-world applications.Fifteen years ago, nonlinear system identification was a field of several ad-hoc approaches, each applicable only to a very restricted class of systems. With the advent of neural networks, fuzzy models, and modern structure optimization techniques a much wider class of systems can be handled. Although one major characteristic of nonlinear systems is that almost every nonlinear system is unique, tools have been developed that allow the use of the same approach for a broad variety of systems. Certainly, a more problem-specific procedure typically promises superior performance, but from an industrial point of view a good tradeoff between development effort and performance is the decisive criterion for success. This book presents neural networks and fuzzy models together with major classical approaches in a unified framework. The strict distinction between the model architectures on the one hand (Part II) and the techniques for fitting these models to data on the other hand (Part I) tries to overcome the confusing mixture between both that is frequently encountered in the neuro and fuzzy literature. Nonlinear system identification is currently a field of very active research; many new methods will be developed and old methods will be refined. Nevertheless, I am confident that the underlying principles will continue to be valuable in the future.This book offers enough material for a two-semester course on optimization and nonlinear system identification. A higher level one-semester graduate course can focus on Chap. 7, the complete Part II, and Chaps. 17 to 21 in VI Preface Part III. For a more elementary one-semester course without prerequisites in optimization and linear system identification Chaps. 1 to 4 and 16 might be covered, while Chaps. 10, 14, 18, and 21 can be skipped. Alternatively, a course might omit the dynamic systems in Part III and instead emphasize the optimization techniques and nonlinear static modeling treated in Parts I and II. The applications presented in Part IV focus on a certain model architecture, and will convince th...

Object tracking in videos by evolutionary clustering and locally linear neuro-fuzzy models

“…Locally linear neuro-fuzzy (LLNF) systems [35] are RBF like neural networks with each neuron accompanied with a set of (usually Gaussian) fuzzy membership functions and a linear function of inputs. It is locally linear because the fuzzy functions act as selective weighting functions for the linear part, and weigh high for some regions of input space and weigh low for other parts.…”

A new approach to fuzzy estimation of Takagi–Sugeno model and its applications to optimal control for nonlinear systems

“…In order to identify the nonlinear function, we take 6 points uniformly distributed {0, 0.2, 0.4, 0.6, 0.8, 1} in the interval [0,1]. In the original T-S method [37], The X matrix becomes: The error between the original system and the identified fuzzy model is with an error of 5.1409e -007.…”

“…It is well known that for u = 0, the equation has a unique singular point (0,0) which is an unstable focus. Moreover, when u = 0, the system displays a sustained oscillation independent of initial conditions known as limit cycle of amplitude A = 2 and natural frequency (o=\.…”

Adaptive Predictive Control of Three - Tank - System

“…Various discrete linear models are used to describe dynamic behaviour of controlled systems; see for example the overview in [18]. The most widely applied linear dynamic model is the ARX model.…”