Identification of Parametric Models: from Experimental Data (Communications and Control Engineering) - Hardcover

Walter, Eric ; Pronzato, Luc

9783540761198: Identification of Parametric Models: from Experimental Data (Communications and Control Engineering)

Hardcover

ISBN 10: 3540761195 ISBN 13: 9783540761198

Publisher: Springer, 1997

View all copies of this ISBN edition

2 Used

From � 57.57

0 New

The presentation of a coherent methodology for the estimation of the parameters of mathematical models from experimental data is examined in this volume. Many topics are covered including the choice of the structure of the mathematical model, the choice of a performance criterion to compare models, the optimization of this performance criterion, the evaluation of the uncertainty in the estimated parameters, the design of experiments so as to get the most relevant data and the critical analysis of results. There are also several features unique to the work such as an up-to-date presentation of the methodology for testing models for identifiability and distinguishability and a comprehensive treatment of parametric optimization which includes greater consider ation of numerical aspects and which examines recursive and non-recursive methods for linear and nonlinear models.

"synopsis" may belong to another edition of this title.

From the Author

Squeezing parameters out of experimental data
We did our best to make this book useful to anyone who has to squeeze parameters out of experimental data. Among its distinctive features are (1) a presentation of the methodology for testing linear and nonlinear models for identifiability and distinguishability; (2) an emphasis on other criteria than least squares (although least squares are, of course, considered), and the importance of robustness; (3) a detailed treatment of parametric optimization, including much more consideration of numerical aspects than usual (evaluation of the effect of rounding errors, generation of derivatives of the cost function with respect to the parameters by exact numerical methods, global optimization...); recursive and non-recursive methods are both considered, for models linear or nonlinear in their parameters; (4) a description of deterministic and statistical methods for characterizing the uncertainty in the parameters resulting from that in the data; (5) a much more detailed presentation of experiment design than usual.

"About this title" may belong to another edition of this title.