Nonlinear Identification and Control: A Neural Network Approach

Citation

Liu, G.P. (2002), "Nonlinear Identification and Control: A Neural Network Approach", Industrial Robot, Vol. 29 No. 5, pp. 469-470. https://doi.org/10.1108/ir.2002.29.5.469.2

This book is part of the “Advances in Industrial Control” series, which aims to provide rapid dissemination of the latest research in all aspects of industrial control. Nonlinear Identification and Control comprises nine chapters which demonstrate how neural networks can be applied to nonlinear industrial control systems.

Chapter 1 provides an introduction to neural networks and addresses topics including architectures of neural networks, a model of a neuron, learning and approximation, mathematical preliminaries, and applications for neural networks.

The following four chapters present different methods for nonlinear identification using neural networks. Variable neural networks, dynamical system modelling by neural network, stable nonlinear identification, and sequential nonlinear identification, are amongst the topics discussed in chapter 2, Sequential Nonlinear Identification. Recursive Nonlinear Identification is addressed in chapter 3, while Multiobject Nonlinear Identification is discussed in chapter 4. Subjects covered in these sections include nonlinear modelling by VPBF networks, recursive learning of neural networks, multiobject modelling with neural networks, model selection by GA, and a multiobject identification algorithm. Chapter 5, Wavelet Based Nonlinear Identification, presents wavelet networks, identification using fixed and variable wavelet networks, and identification using B‐Spline wavelets.

The four remaining chapters of the book present various techniques for nonlinear control using neural networks. Chapter 6, Nonlinear Adaptive Neural Control, discusses adaptive neural control and an adaptation algorithm with variable networks, while Nonlinear Predictive Neural Control is addressed in chapter 7. Topics presented in this section include nonlinear neural predictors, on‐line learning of neural predictors, and sequential predictive neural control. Chapters 8 and 9 address Variable Structure Neural Control, and a Neural Control Application to Combustion Processes, respectively. Topics discussed in these sections include generalised variable structure neural control, recursive learning for variable structure control, a model of combustion dynamics, an output predictor and controller, and active control of a simulated and an experimental combustor.

Nonlinear Identification and Control is a well written reference text with each chapter concluding with examples and a summary. This book will be of interest to academics, postgraduate students and industrial engineers.