Robustness in Feedback Systems.
Abstract
The research in this thesis is motivated by the basic question 'What can and cannot be accomplished by feedback control?' In particular, three basic issues are addressed: a) Determining which types of controllers are optimal for certain classes of control problems; b) Investigating the absolute limitations of feedback control for multiobjectives; and c) Developing efficient methods for synthesizing robustly stabilizing controllers for families of plants featuring block-structed uncertainty. With regard to these issues the principal contributions are: as follows. First, it is shown that for the problem of robustly stabilizing a family of plants featuring dynamic uncertainty, linear time-invariant controllers perform as well as arbitrary nonlinear time-varying controllers. Second, a new controller synthesis procedure called residue iteration is developed for synthesizing robustly stabilizing controllers for families of plants featuring block-structured uncertainty. This method is simpler and numerically more attractive than any previously existing technique. Finally, an algorithm is presented which enables one to compute an absolute upper bound on the performance levels attainable in multiobjective H at infinity-optimization problems.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1987
- Accession Number
- ADA182830
Entities
People
- Thomas L. Ting
Organizations
- University of Illinois Urbana–Champaign