• Media type: E-Article
  • Title: Robust nonfragile observer-based H2/H∞ controller
  • Contributor: Oveisi, Atta; Nestorović, Tamara
  • Published: SAGE Publications, 2018
  • Published in: Journal of Vibration and Control, 24 (2018) 4, Seite 722-738
  • Language: English
  • DOI: 10.1177/1077546316651548
  • ISSN: 1077-5463; 1741-2986
  • Origination:
  • Footnote:
  • Description: A robust nonfragile observer-based controller for a linear time-invariant system with structured uncertainty is introduced. The [Formula: see text] robust stability of the closed-loop system is guaranteed by use of the Lyapunov theorem in the presence of undesirable disturbance. For the sake of addressing the fragility problem, independent sets of time-dependent gain-uncertainties are assumed to be existing for the controller and the observer elements. In order to satisfy the arbitrary H2-normed constraints for the control system and to enable automatic determination of the optimal [Formula: see text] bound of the performance functions in disturbance rejection control, additional necessary and sufficient conditions are presented in a linear matrix equality/inequality framework. The [Formula: see text] observer-based controller is then transformed into an optimization problem of coupled set of linear matrix equalities/inequality that can be solved iteratively by use of numerical software such as Scilab. Finally, concerning the evaluation of the performance of the controller, the control system is implemented in real time on a mechanical system, aiming at vibration suppression. The plant under study is a multi-input single-output clamped-free piezo-laminated smart beam. The nominal mathematical reduced-order model of the beam with piezo-actuators is used to design the proposed controller and then the control system is implemented experimentally on the full-order real-time system. The results show that the closed-loop system has a robust performance in rejecting the disturbance in the presence of the structured uncertainty and in the presence of the unmodeled dynamics.