Output feedback stabilization of systems with uncertainty (Q1395029)

The aim of this paper is to study the output feedback stabilization for an uncertain controlled system of a high relative order having the form \[ \dot x(t)=Ax(t)+b(u(t)+\psi(t,x)),\quad t\geq 0, \] with the uncertainty \(\psi\in {\mathcal K}(t,x)\) and the observed output \[ y(t)=ax(t), \] where \({\mathcal K}:\mathbb R_+\times \mathbb R^n\to\mathbb R\) is a many-valued mapping, piecewise-continuous in \(t\) and continuous in \(x\), with linearly bounded growth. A method is proposed which partitions the original problem of \(\gamma\)-exponential stabilization of high-\(r\) systems into two: a problem of \(\gamma\)-exponential invariance of an \(r=1\) system and a problem of dynamic uncertainty correction. This fact leads to a natural separation of the control goals for the two subsystems: for one subsystem the stabilization problem is solved while the control of the other subsystem is aimed to satisfy the adjustment conditions, which are stated in the form of a functional relationship. Consequently, this approach consists of three stages: problem reduction, dynamic correction of the uncertainty, and ensuring \(\gamma\)-exponential invariance of the \(r=1\) system.