Decoupling control (Q1847336)

A given multi-variable plant usually has cross-couplings or interaction between its variables, i.e. one input variable may affect all the output variables. The aim of the decoupling control is to design a multivariable controller which can decouple the plant, that is, the resulting control system has no more couplings between the desired reference variables and the plant output variables. Therefore, decoupling is very useful in practical applications, since it simplifies the control system design. The study of the decoupling of linear time-invariant systems has received considerable attention in both control theory and industrial practice for several decades. As the decoupling control is nowadays well developed, at least for nominal design, the author meets the challenge of writing a book that gives comprehensive and in-depth coverage of this topic. This book is the first monograph which is devoted to decoupling control, and many recent results have been obtained by the author in the long series of papers indicated in the references. Each paragraph contains examples that illustrate the theory, and each chapter ends with extended references to the work that has been presented. The text is structured in 10 chapters. Chapter 1 introduces the history and philosophy of decoupling control. Chapter 2 starts with some basic concepts of linear systems, such as the solution of state space systems, controllability and stabilizability. An introduction to the known polynomial and rational matrix theory is approached in the sequel by known results from Rosenbrock (1974), Callier and Desoer (1982), Wolovich (1978) and Kucera (1979). Then the model reduction problem for single input single output (siso) systems is studied by using two different approaches: the recursive least squares approach, where stability can be taken care of but cannot be guaranteed, and the step response construction, where more steps are needed but stability is preserved. This chapter ends with reference to the known conversion methods between continuous and discrete time systems (i.e. forward, backward and bilinear rule, and zero-order hold), while a new frequency response fitting approach has been introduced by the author with better results than the other known ones, as concerning the approximation error. In Chapter 3 two useful and easily implemented theorems for the internal stability of interconnected systems are established. The first theorem is for scalar systems, whereas the second one is for vector signals. The problem of finding a proper controller in a unity feedback system by an algebraic approach is adressed in the sequel, while for the case of constant controllers the multivariable Nyquist and Gershgorin's theorems are given. Finally, a presentation of structured and unstructured uncertainties and their effects on control systems is given. The importance of limited phase uncertainty for the robust stability criterion, in addition to the existing magnitude information, is also mentioned. Chapter 4, based on the algebraic approach of Falb and Wolovich (1967) and Wolovich and Falb (1969), presents necessary and sufficient conditions for the static and dynamic decoupling of state space systems by using a state feedback controller. Furthermore, it shows the way we can change the poles of a decoupled system by applying a state feedback. However, the zeros of the decoupled system remain unchanged. In Chapter 5 necessary and sufficient conditions for the solvability of the general decoupling problem by unity output feedback compensation are established. The block decoupling problem is studied first for square plants and then for nonsquare plants, while a unified solution is presented in the sequel. A parametrization of all the decoupling compensators is achieved. A number of necessary tools for the derivation of the above results, such as skew primeness, strict adjoints, stability factorization and their extension to non-square plants, are also presented. In Chapter 6 and 7 new approaches to the decoupling problem for plants with time delays are addressed and decoupling methodologies with necessary theoretical supports, as well as the controller design details for possible practical applications, are presented. Chapter 6 considers a conventional unity output feedback configuration, whereas Chapter 7 exclusively deals with time delay compensation schemes. What is really needed in practice is not exact decoupling but, instead, to limit the loop interaction to an acceptable level. For that reason, Chapter 8 proposes the concept of near decoupling and develops an approach to near-decoupling controller design for both nominal and uncertain systems, by using state or dynamic output feedback. The aim of near decoupling is to design a decoupling law that makes the magnitude of off-diagonal elements in the transfer matrix of a system as small as possible, instead of zero, under the existence of parameter perturbations, while ensuring that the magnitude of diagonal elements in the transfer matrix are larger than a given number. Chapter 9 and 10 deal with decoupling disturbances from the plant output in dynamic and asymptotic senses respectively. The aim of the dynamic disturbance decoupling is to find a control scheme such that in the resulting control system: a) the transfer matrix function from the disturbance (measured or unmeasured) to the controlled output is zero for all frequencies and b) the resulting system is stable. In cases where the disturbances are measured, the feedforwarded scheme is employed, whereas in the most frequent case of unmeasured disturbances, a disturbance observer is introduced that acts as an add-on mechanism to the conventional feedback system. The aim of the static or asymptotic disturbance decoupling is to decouple the disturbance from the plant output at the steady-state only. The main focus for that case is when we have unmeasurable disturbances, since for measurable disturbances the asymptotic disturbance decoupling problem is straightforward. Finally, the emphasis of future research, according to the author, should be laid on robust decoupling and, in particular, practical and stable design techniques which can achieve the specified coupling constraints under the given plant uncertainty.