Identification of Dynamic Systems: An Introduction with Applications
Precise dynamic models of processes are required for many applications, ranging from control engineering to the natural sciences and economics. Frequently, such precise models cannot be derived using theoretical considerations alone. Therefore, they must be determined experimentally. This book treats the determination of dynamic models based on measurements taken at the process, which is known as system identification or process identification. Both offline and online methods are presented, i.e. methods that post-process the measured data as well as methods that provide models during the measurement. The book is theory-oriented and application-oriented and most methods covered have been used successfully in practical applications for many different processes. Illustrative examples in this book with real measured data range from hydraulic and electric actuators up to combustion engines. Real experimental data is also provided on the Springer webpage, allowing readers to gather their first experience with the methods presented in this book. Among others, the book covers the following subjects: determination of the non-parametric frequency response, (fast) Fourier transform, correlation analysis, parameter estimation with a focus on the method of Least Squares and modifications, identification of time-variant processes, identification in closed-loop, identification of continuous time processes, and subspace methods. Some methods for nonlinear system identification are also considered, such as the Extended Kalman filter and neural networks. The different methods are compared by using a real three-mass oscillator process, a model of a drive train. For many identification methods, hints for the practical implementation and application are provided. The book is intended to meet the needs of students and practicing engineers working in research and development, design and manufacturing.
Why Read This Book
You should read this book if you need a compact but rigorous introduction to system identification that balances theory and practical application. You will learn how to build, validate, and use dynamic models from measured data using both batch and recursive methods, with many engineering examples.
Who Will Benefit
Graduate students, control and signal-processing engineers, and practitioners who need to estimate dynamic process models for control, prediction or fault diagnosis.
Level: Advanced — Prerequisites: Undergraduate-level signals & systems and linear systems theory, basic probability/statistics, linear algebra, and familiarity with MATLAB (recommended for examples).
Key Takeaways
- Formulate system identification problems and choose suitable model structures (ARX/ARMAX/OE/Box-Jenkins, state-space).
- Apply least-squares and prediction-error estimation methods to obtain consistent parameter estimates.
- Use frequency-domain and nonparametric methods for spectral estimation and model insight.
- Employ subspace and state-space identification techniques to derive modern reduced-order models.
- Design experiments and validate models using residual analysis and statistical tests.
- Implement recursive/online identification algorithms for adaptive modeling in real-time applications.
Topics Covered
- 1. Introduction and Motivation
- 2. Problem Formulation and Model Structures
- 3. Least-Squares and Prediction Error Methods
- 4. Frequency-Domain and Nonparametric Estimation
- 5. Parametric Model Classes (ARX, ARMAX, OE, Box-Jenkins)
- 6. State-Space and Subspace Identification
- 7. Recursive and Online Identification Methods
- 8. Identification in Closed-Loop and Practical Issues
- 9. Experiment Design and Signal Choice
- 10. Model Validation, Residual Analysis and Statistical Tests
- 11. Nonlinear System Identification (overview and methods)
- 12. Applications and Case Studies
- Appendices: Mathematical Tools and Implementation Notes
Languages, Platforms & Tools
How It Compares
Comparable to Ljung's System Identification (more exhaustive and theory-heavy), Isermann is shorter and more application-oriented; also overlaps with Pintelon & Schoukens but with a stronger control/practical slant.
