Lecture Notes in Control and Information Sciences
2 primary works
Book 138
Robust Controller Design Using Normalized Coprime Factor Plant Descriptions
by Duncan C. McFarlane and Keith Glover
Published 11 December 1989
This monograph concerns the design of feedback controllers for multivariable linear dynamic systems. The particular approach is to consider a coprime factor description of the plant's transfer function and to represent a family of systems by perturbing the numerator and denominator. The design of controllers to robustly stabilize such a family is posed as an H? optimization problem and some explicit solutions are obtained. Similarly, procedures for reduced order modelling and controller design are derived. Finally, the results are exploited to give a systematic loop shaping control system design procedure that is assessed on several aerospace examples. The book will be appropriate for advanced undergraduate or graduate classes requiring only a first course in state-space methods. It also gives a good introduction to multivariable control and the use of H? methods.
Book 146
This monograph is concerned with the design of feedback controllers for linear multivariable systems, which are robust to system uncertainty. System uncertainty can be realistically represented by including perturbations with bounded H?-norm, and this is the approach taken here. For a given H?-norm bound, there is a family of robustly stabilizing controllers, and the central question in this book is which of these controllers to choose. One choice to take is that which minimizes the enthropy of the resulting closed loop transfer function, and the derivation and properties of this solution occupies most of this monograph. Explicit formulae are obtained for the minimum enthropy solution, which is a precisely defined compromise between the Linear Quadratic Gaussian optimal solution and the H?-optimal solution. The book will be appropriate for graduate classes requiring only a first course in state-space methods, and some elementary knowledge of H? control and Linear Quadratic Gaussian control.