Title :
On the robustness of MEOP design versus asymptotic LQG synthesis
Author :
Cheung, Man-Fung ; Yurkovich, Stephen
Author_Institution :
Dept. of Electr. Eng., Ohio State Univ., Columbus, OH, USA
Abstract :
An analysis is presented of generalized LQG (linear-quadratic Gaussian)-type algorithms with regard to robustness characteristics. Comparisons are made of the maximum entropy/optimal projection (MEOP) methodology versus the linear quadratic Gaussian with loop transfer recovery (LQG/LTR) technique and the parametric robust linear quadratic Gaussian synthesis procedure. The performance of a reduced-order direct controller design is compared with the full-order designs on a example representative of typical flexible-structure dynamics.<>
Keywords :
control system synthesis; optimal control; stability; LQG/LTR; control system synthesis; flexible-structure dynamics; linear quadratic Gaussian; loop transfer recovery; maximum entropy/optimal projection; optimal control; parametric robust linear quadratic Gaussian synthesis; reduced-order direct controller; stability; Boundary conditions; Control system synthesis; Design methodology; Equations; Flexible structures; GSM; Linear systems; Regulators; Robust control; Robustness;
Journal_Title :
Automatic Control, IEEE Transactions on
