×

Design of robust observation schemes for uncertain large scale systems. (English) Zbl 0892.93007

The author presents a robust observer design for uncertain large-scale systems with nonlinear crosscouplings. Uncertainty is represented by nonlinear functions satisfying a kind of matching condition and bounded in norm by know scalar functions of output and input variables. The proposed observer is decentralized and nonlinear and takes into account functional properties of the uncertainty bounds. The gain matrices in the observer are given in terms of the solutions of constrained Riccati equations. The nonlinear crosscouplings between subsystems are decomposed into matched and mismatched parts. The mismatched part may result in an iterative search for Lyapunov functions used to design a nonlinear part of the observer. This part has a generalized sign form used (e.g. [Z. Qu, Automatica 29, 985-998 (1993; Zbl 0776.93041)], [Y. H. Chen, Int. J. Control 46, 1979-1992 (1987; Zbl 0634.93005)]) also in robust stabilization design procedures by many authors. The author proposes also a search algorithm for choosing the required parameters and proves the exponential convergence of the observation error. Moreover, new conditions the solvability of the proposed constrained Riccati equations are presented.

MSC:

93A15 Large-scale systems
93B07 Observability
93D30 Lyapunov and storage functions
93B35 Sensitivity (robustness)
93C10 Nonlinear systems in control theory
PDFBibTeX XMLCite
Full Text: DOI