Fuzzy model identification for control, J. Abonyi, Birkhauser, Boston, 2003, x+273 pages
Bruce E. Postlethwaite · International Journal of Robust and Nonlinear Control · 2004
In the last decade model-based control has become increasingly widespread in the process industries.Despite the fact that the processes being controlled are universally nonlinear, the model-based schemes in current industrial use almost all use linear models.Fuzzy modelling offers one way of introducing nonlinear models into control schemes, and has been the subject of research by many groups around the world.This book is principally a description of the work carried out by one of these groups.The book is divided into five chapters, with over 130 figures, 280 references and many examples.Chapter 1 is an introduction to the book.The first section of this chapter is concerned with 'Fuzzy modelling with the use of prior knowledge'.This seems a bit strange as the first section, but is really an indication of the scope of the rest of the book}it is very much directed at a specific area of fuzzy modelling.The author doesn't try to conceal this, indeed he makes it clear on the first page, but I can't help feeling that the title of the book indicates something more general.Most of the rest of the chapter is concerned with a discussion of the philosophy of modelling (e.g.white box vs black box vs fuzzy model) as related to the concerns of the author.This is a wellwritten and thoughtful section and is a very good introduction into the ways in which a priori knowledge can be mixed with data driven modelling approaches.There are almost no mathematical equations in this section}everything is explained in text and diagrams.I find this refreshing compared to introductions where the authors present huge blocks of complicated equations.Abonyi has obviously taken time to reflect on his modelling approach and is capable of expressing his thoughts clearly and with authority.The chapter rounds off with a very