Coercive ISS-Lyapunov functionals for regular infinite-dimensional systems and applications
Swann Marx
TL;DR
This work tackles the lack of coercivity in ISS-Lyapunov functionals for regular infinite-dimensional systems with outputs. It shows that if the output operator yields exact observability, a noncoercive ISS Lyapunov functional can be augmented by the observability Gramian to become coercive, producing a Lyapunov inequality that includes an explicit output term. The approach relies on regular linear systems theory and constructs the coercive functional via a Gramian-based additive term, enabling applications to singular perturbation analyses and output regulation, with the KdV equation serving as a nontrivial demonstration. The results unify ISS methodology with exactly observable outputs and offer pathways to strictification through observer designs, potentially impacting controller and observer synthesis for distributed-parameter systems.
Abstract
This paper proposes the construction of a coercive ISS-Lyapunov functional for linear regular infinite-dimensional system. Indeed, as already known, Lyapunov functionals for infinite-dimensional systems might be not coercive. Under the assumption that there exists an exactly observable output, we are able to make coercive a Lyapunov functional which is not coercive under additional regularity assumption. We discuss also about the potential applications of such a Lyapunov functional in singular perturbation theory and output regulation. The results are illustrated on a non-trivial equation, namely, the Korteweg-de Vries equation.
