Stability analysis of uncertain systems via matrix-valued Liapunov functions (Q2568097)

The author studies the following uncertain system of differential equations \[ \frac{dx(t)}{dt}=f(t,x,\alpha),\quad x(t_0)=x_0,\tag{1} \] with \(x(t)\in \mathbb{R}^n, t\in T_0=[t_0, +\infty), t_0\in T_i, \) and \(T_i\subseteq \mathbb{R}\), and \(f\in C(T_0\times \mathbb{R}^n\times \mathbb{R}^d, \mathbb{R}^n).\) The parameter \(\alpha\in \Omega\subseteq R^d\), where \(\Omega\) is a compact set, is the ``uncertainty'' parameter of system (1). In system (1), as described in the work of \textit{Y.-H. Chen} [Dyn. Control 6, 131--142 (1996; Zbl 0850.93731)], the parameter \(\alpha\) can: (a) present an uncertain value of a physical parameter of the system or an estimate on the external perturbation; (b) be a function mapping \(\mathbb{R}\) into \(\mathbb{R}^d\) and presenting an uncertainly measured value of the input effects of one subsystem of the other; (c) be a function mapping \(T_0\times \mathbb{R}^n\) into \(\mathbb{R}^d\) and presenting nonlinear elements of the mechanical system which are difficult for exact measurings; (d) be an index which shows the existence of some uncertainties in the system; and (e) be a combination of characteristics (a)--(c). A new approach is given to investigate the qualitative motion properties of the uncertain system (1). Sufficient conditions are given to guarantee the motions of the uncertain system (1) possess a certain type of stability by using a generalized direct Lyapunov method based on matrix-valued auxiliary functions.