Lyapunov-like conditions of forward invariance and boundedness for a class of unstable systems (Q2848586)

Motivated by some important problems in system theory (synchronization, adaptive control), the authors provide new tools for the stability analysis of systems of the form NEWLINE\[NEWLINE\begin{aligned} & \dot x = f(x,\lambda,t), \\ & \dot \lambda = g(x,\lambda,t), \end{aligned}NEWLINE\]NEWLINE where \(x\in \mathbb R^n\), \(\lambda\in\mathbb R\); \(f\) and \(g\) satisfy some usual regularity conditions, and \(g\) is of constant sign. The origin is assumed to be an equilibrium point. Note that because of the assumption about the sign of \(g\), the origin may be not stable (in the sense of Lyapunov). Moreover, it is assumed that for the system NEWLINE\[NEWLINE \dot x = f(x,0,t)NEWLINE\]NEWLINE the set of initial states attracted by the origin is of positive measure. The authors propose an extension of the classical method based on the use of a Lyapunov-like function, in order to establish the existence of forward invariant sets (not necessarily neighborhoods of the origin) and sets of initial conditions whose corresponding solutions converge to the origin. The same approach can be used to identify sets of initial states whose corresponding solutions escape a given neighborhood of the origin. Examples and applications are illustrated.