System Identification and Modeling

Author: Pan Lideng
Publisher:
Publish Date: 2004-01-01
Features: This book progresses step by step, clarifying the relationship between theory and algorithms, and includes some programs to facilitate readers' understanding, mastery, and practical application and programming. It can serve as a textbook for master's students in automation, systems engineering, and testing technology, or as a selective course textbook for senior undergraduate students in the same fields. It can also be referenced by teachers, researchers, and engineering technicians in these fields. System identification and modeling are theories and methods for establishing mathematical models of production processes using input and output data from observed processes. They provide the basis for improving control system quality, designing advanced control systems, and achieving optimal control. This book takes the least squares theory as its main thread, introducing various least squares methods such as augmented least squares, generalized least squares, multistep least squares, auxiliary variable methods, maximum likelihood methods, Kalman filtering, model reference adaptive methods, stochastic direct search methods, stochastic approximation methods, multivariable system identification methods, closed-loop system identification methods, and small-sample system identification methods. It also studies identification methods for various model orders and their industrial applications. Particularly, the stochastic direct search method, the main model decomposition submodel method in multivariable system identification, and the numerical state-space subspace method (N4SID) are the highlights of this book and play a significant role in modeling.