Identifiability and Observability via decoupled variables: Application to a malaria intra-host model
Abstract
For a class of systems for which identifiability is difficult to be proved in the original set of coordinates, we show that when there exists a change of variables which decouples some variables i.e., whose dynamics is independent of the parameters, identifiability can then be assessed without having to compute derivatives. Moreover, we show that this allows to consider an observer to reconstruct the state of the system without having to estimate the unknown parameters. The effectiveness of this approach is illustrated on an intra-host model of malaria infection.
Origin : Files produced by the author(s)