Observability to the identifiability of hybrid dynamical system

Abstract

Hybrid complex systems are heterogeneous dynamical systems whose behavior can be defined by interacting continuously and by having discrete changes in their dynamics. Therefore, this class of systems presents commutations which can be autonomous or driven by some external events. So, the identification of this class of systems is a challenging problem that involves estimation of both parameters of the sub-models, and the partitioning of states, data and input space. In this paper, an approach to modeling this kind of system is introduced and analyzed for observability and identifiability properties.