2 internautes sur 2 ont trouvé ce commentaire utile
- Publié sur Amazon.com
This book covers most aspects related to system identification: linear, nonlinear, discrete-time, continuous-time, state-space methods, frequency domain and time domain methods, closed-loop, open-loop etc. In this respect it is a far more comprehensive look at system identification than most of the textbooks available in the English language. The presentation of the material does not follow Ljung's notational system -- which is good news to many readers--, and is more readable. The book is quite self-contained in the sense that there is a lot of fundamental material on stochastic process and linear system theory (e.g. state-space representation) that is usually assumed known in most books. The book at times is not as clear as one would wish, although I cannot say this is a major problem.
If you are considering this as a textbook, then you should know that it comes with very few examples that aim at illustrating how to implement the methods. The provided examples, in general, compare the performance of identification methods by means of plots, hence there is no real help to the reader who would like to see how to implement a certain method. The last chapter of the book, which is the longest in the book, deals with a nice variety of study cases but they are more illustrative than pedagogical. Also the end-of-chapter problems are usually of a general character e.g. "Describe an advantageous test signal sequence". Therefore whoever adopts this book as a course textbook will have to do extra work in providing examples and problems. Related to this, the book offers various sets of data that can be used to test methods and algorithms.
Special mention is due to the references used by the authors. There are three aspects which I would like to comment on. On the positive side is the care the authors have paid to the historical development of the field. On most of the topics there are several early references that help the reader to grasp a sense of the time-frame in which the methods were developed. There is a relatively large number of references in German. This is a positive point (for those who have access to such references) and reveals the authors' care in quoting their fellow countrymen. On the negative side is the absence of some fundamental, almost "classic", references about some of the topics in the second half of the book. Since most of the current research topics are concentrated there, the book is of little help in pointing to important references.