A stabilization algorithm for a class of uncertain linear systems (Q1089299): Difference between revisions

Necessary and sufficient conditions are given for the quadratic stabilization of the uncertain control system \[ \dot x(t)=(A+DF(t)E)x(t)+Bu(t),\quad F^ T(t)F(t)\leq I, \] \(x\in {\mathfrak R}^ n\), \(u\in {\mathfrak R}^ m\), and \(F\in {\mathfrak R}^{p\times q}\). The form of the stabilizing feedback control is \(u^*(t)=-R^{-1}B^ TPx(t)\), where R and P are positive definite matrices of appropriate dimension, and P solves an algebraic Riccati equation. It is demonstrated that with the notion of ''overbounding'', it is possible to use the control law \(u^*(t)\) to stabilize larger classes of uncertain linear systems.