Advanced Control Law Tuning and Performance Assessment
Abstract
It is evident that military applications for 21st century will be highly complex, multivariable systems. Designing optimal controllers for such applications will require the use of mathematical models which describe the complexity of the underlying processes as accurately as possible. For robustness it is essential that these models include the uncertainties in the estimated process dynamics and trajectories. Controllers resulting from optimal control strategies (like LQG, H2, H(infinity)) will usually be of very high order and that can cause implementation and computational problems. Some of these problems can be overcome by using controllers that are of lower order and restricted structure. The subject of restricted structure controller design and performance assessment is relatively new and enables the expected performance to be assessed against a much more realistic criterion. That is, the performance figures take into account the limitations of the existing control system structure, and hence provide a more accurate measure of the possible performance improvement. However, the design and/or tuning of restricted-structure controllers in order to provide the performance comparable with full-order solutions is still a very contentious issue. Added to this is the need to achieve robust properties and performance specifications required by military applications. Very few methods for designing restricted-structure controllers exist that allow the performance and robustness objectives to be combined into one relatively simple optimization problem. This lecture presents an LQG/H2-based method that tackles the above mentioned issues.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 2006
- Accession Number
- ADA470887
Entities
People
- Michael J. Grimble
- Pawel Majecki
Organizations
- University of Strathclyde