Finite-time and fixed-time input-to-state stability: explicit and implicit approaches (Q826761)

Given the nonlinear system \(\dot{x}(t) = f(x(t), d(t))\), where \(x\) and \(d\) denote the state and the input, respectively, input-to-state stability (ISS) means that the state satisfies the following estimate; \(\|x(t)\| \leq \beta(\|x_0\|,t) + \nu(\|d\|_{[0,t]})\) for all \(t \geq 0\). Here \(\nu\) is increasing, \(\beta\) is increasing in the first argument, but decreasing in the second, and \(\|d\|_{[0,t]}\) is the supremum of \(\|d\|\) on the interval \([0,t]\). When \(\beta(r,t)\) can be chosen be be zero for \(t\) sufficiently large, then we speak of finite-time ISS. When this time can be chosen independently of \(r\), we say that the system is fixed-time ISS. The aim of the paper is to relate these stability properties to the existence of Lyapunov functions. Since the input appears in the differential equation, it appears in the conditions for the Lyapunov function as well. \(V\) is a ISS Lyapunov function if \(\alpha_1(\|x\|) \leq V(x) \leq \alpha_2(\|x\|)\) and \(\dot{V}(x) \leq \delta(\|d\|)-\xi(\|x\|)\), where the functions \(\alpha_1,\alpha_2,\delta\) and \(\xi\) are strictly increasing, unbounded functions. It is known that ISS is equivalent to the existence of an ISS Lyapunov function, and the authors extend that result to the other ISS stability notions. It is shown that a sufficient condition for finite-time or fixed-time ISS is the existence of an ISS Lyapunov function with some additional properties. Conversely, if \((\|f(x,d)- f(x,0)\|)/\|d\|\) can be bounded independently of \(d\), then finite-time ISS implies the existence of a corresponding ISS Lyapunov function. By extended the class of Lyapunov functions to those implicitly defined, they extend these theorems.