Stabilization of infinite-dimensional second-order systems by adaptive PI-controllers

From MaRDI portal





Consider the second order abstract differential equation \(\ddot{x}(t)+ A x(t) = B u(t)\), \(y(t)= C\dot{x}(t)\), with \(A\) selfadjoint and non-negative, and \(B\) and \(C\) unbounded operators. The aim of this paper is to stabilize this systems without explicit knowledge of the system. Since the systems has all its poles/eigenvalues on the imaginary axis, and since \(B\) and \(C\) are unbounded, the standard adaptive controller based on high gain cannot be applied. Under some additional technical assumptions the authors shows that the adaptive controller \(u(t) = -K(t)\int_0^t y(\tau) d\tau - k(t) y(t)\), with \(K(t)=\gamma y(t) \int_0^t y(\tau) d\tau + \kappa\) and \(\dot{k}(t)= r y(t)^2\), where \(\gamma,\kappa,k(0), r >0\) stabilizes the system. This is shown by proving that the energy converges to zero. The theory is illustrated by means of two elastic string examples.











This page was built for publication: Stabilization of infinite-dimensional second-order systems by adaptive PI-controllers

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2731605)