An introduction to the emerging dominant paradigm in control design
Model-based approaches to control systems design have long dominated the control systems design methodologies. However, most models require substantial prior or assumed information regarding the plant’s structure and internal dynamics. The data-driven paradigm in control systems design, which has proliferated rapidly in recent decades, requires only observed input-output data from plants, making it more flexible and broadly applicable.
An Introduction to Data-Driven Control Systems provides a foundational overview of data-driven control systems methodologies. It presents key concepts and theories in an accessible way, without the need for the complex mathematics typically associated with technical publications in the field, and raises the important issues involved in applying these approaches. The result is a highly readable introduction to what promises to become the dominant control systems design paradigm.
Readers will also find: - An overview of philosophical-historical issues accompanying the emergence of data-driven control systems - Design analysis of several conventional data-driven control systems design methodologies - Algorithms and simulation results, with numerous examples, to facilitate the implementation of methods
An Introduction to Data-Driven Control Systems is ideal for students and researchers in control theory or any other research area related to plant design and production.
Table of Contents
Preface xi
Acknowledgements xv
List of Acronyms xvii
1 Introduction 1
1.1 Model-Based Control System Design Approach 1
1.2 Data-driven Control System Design Approach 5
1.3 Data-Driven Control Schemes 10
1.4 Outline of the Book 25
2 Philosophical Perspectives of the Paradigm Shift in Control Systems Design and the Re-Emergence of Data-Driven Control 35
2.1 Introduction 35
2.2 Background Materials 36
2.3 Paradigm Shifts in Control Systems Design 44
2.4 Uncertainty Combat Paradigm 54
2.5 The Paradigm Shift Towards Data-driven Control Methodologies 61
2.6 Conclusions 68
3 Unfalsified Adaptive Switching Supervisory Control 73
3.1 Introduction 73
3.2 A Philosophical Perspective 75
3.3 Principles of the Unfalsified Adaptive Switching Control 77
3.4 The Non-Minimum Phase Controller 87
3.5 The DAL Phenomena 88
3.6 Performance Improvement Techniques 91
3.7 Increasing Cost Level Algorithms in UASC 95
3.8 Time-varying Systems in the UASC 101
3.9 Conclusion 104
4 Multi-Model Unfalsified Adaptive Switching Supervisory Control 111
4.1 Introduction 111
4.2 The Multi-Model Adaptive Control 113
4.3 Principles of the Multi-Model Unfalsified Adaptive Switching Control 116
4.4 Performance Enhancement Techniques in the MMUASC 126
4.5 Input-constrained Multi-Model Unfalsified Switching Control Design 129
4.6 Conclusion 147
5 Data-Driven Control System Design Based on the Virtual Reference Feedback Tuning Approach 155
5.1 Introduction 155
5.2 The Basic VRFT Methodology 156
5.3 The Measurement Noise Effect 163
5.4 The Non-Minimum Phase Plants Challenge in the VRFT Design Approach 166
5.5 Extensions of the VRTF Methodology to Multivariable Plants 171
5.6 Optimal Reference Model Selection in the VRFT Methodology 177
5.7 Closed-loop Stability of the VRFT-Based Data-Driven Control Systems 183
5.8 Conclusions 187
6 The Simultaneous Perturbation Stochastic Approximation-Based Data-Driven Control Design 193
6.1 Introduction 193
6.2 The Essentials of the SPSA Algorithm 195
6.3 Data-Driven Control Design Based on the SPSA Algorithm 201
6.4 A Case Study: Data-Driven Control of Under-actuated Systems 205
6.5 Conclusions 212
7 Data-driven Control System Design Based on the Fundamental Lemma 217
7.1 Introduction 217
7.2 The Fundamental Lemma 218
7.3 System Representation and Identification of LTI Systems 222
7.4 Data-driven State-feedback Stabilisation 225
7.5 Robust Data-driven State-feedback Stabilisation 228
7.6 Data-driven Predictive Control 233
7.7 Conclusion 247
8 Koopman Theory and Data-driven Control System Design of Nonlinear Systems 253
8.1 Introduction 253
8.2 Fundamentals of Koopman Theory for Data-driven Control System Design 254
8.3 Koopman-based Data-driven Control of Nonlinear Systems 269
8.4 A Case Study: Data-driven Koopman Predictive Control of the ACUREX Parabolic Solar Collector Field 281
8.5 Conclusion 288
9 Model-free Adaptive Control Design 293
9.1 Introduction 293
9.2 The Dynamic Linearisation Methodologies 295
9.3 Extensions of the Dynamic Linearisation Methodologies to Multivariable Plants 302
9.4 Design of Model-free Adaptive Control Systems for Unknown Nonlinear Plants 304
9.5 Extensions of the Model-free Adaptive Control Methodologies to Multivariable Plants 314
9.6 A Combined MFAC-SPSA Data-driven Control Strategy 330
9.7 Conclusions 337
Problems 338
References 339
Appendix 341
A Norms 341
B Lyapunov Equation 343
C Incremental Stability 343
D Switching and the Dwell-time 344
E Inverse Moments 346
F Least Squares Estimation 349
G Linear Matrix Inequalities 351
H Linear Fractional Transformations 353
References 355
Index 357
