Predictive control with constraints (Q2735974)

The book, consisting of a Preface, 10 Chapters, References, and 3 Appendices, presents model-based predictive control (PC) with constraints. NEWLINENEWLINENEWLINEChapter 1 introduces the main concepts involved in PC such as internal model, reference trajectory, on-line optimization. Chapter 2 presents a standard formulation of PC with constraints for a multivariable plant in a state-space setting. The concepts of an independent model, offset-free tracking and stabilization of an unstable model are discussed in the context of state observers. Chapter 3 deals with solving PC problems under constraints mainly by using active set and interior point methods. The structure of the resulting controller is discussed. Softening the constraints is described as a way to solve the situation when the optimization problem becomes infeasible. Chapter 4 presents the formulation of PC based on step response and transfer function models, and shows their relation to the state-space setting. Chapter 5 presents other formulations of PC such as the usage of feedforward or non-quadratic costs, predictive functional control, and continuous-time PC. Chapter 6 presents methods guaranteeing closed-loop nominal system stability. Two principal methods are presented to support this: terminal constraints and infinite horizon. Further, fake algebraic Riccati equations and the Youla parametrization are also introduced. Chapter 7 deals with tuning of controller parameters to optimize the prediction horizon, the time constant of the reference trajectory, and weights in the criteria. Chapter 8 considers PC robust to parameter uncertainties. Some recent research proposals based on a tuning procedure of Lee and Yu, the LQG/LTR tuning procedure, the LMI approach, as well as robust feasibility are introduced. This problem is still under research. Chapter 9 presents two case studies. Chapter 10 considers some perspectives of PC such as the possibilities of constraint management, the use of nonlinear models, and the potential of PC for fault-tolerant control. The appendices provide details of some software issues: leading commercial products, software used in the book, and MATLAB's MPC Toolbox. NEWLINENEWLINENEWLINEThe book provides a systematic and comprehensive course on predictive control with mini-tutorials on special topics, numerous examples and exercises. This well-readable and self-contained textbook which promotes the use of MATLAB is primarily addressed to graduate students as well as practicing engineers.