The field of optimal input design (OID) considers the problem of finding an input signal that leads to the most informative experiment, given some prior knowledge about the system, while respecting the physical limitations of the measurement setup. Mathematically, the OID problem can be formulated as an optimization problem in which a scalar measure of the Fisher information matrix is maximized with respect to the parametrization of the input sequence. In this study, the OID is computed for a Wiener model that consists of a combination of a linear time invariant system, followed by power nonlinearity of known order, while the class of inputs is restricted to zero-order-hold signals with fixed range or fixed total energy. To facilitate the computation of the Fisher information matrix, the linear system is parameterized with its poles and residues instead of the more classical rational form representation. Due to the nonlinearity of the system, the resulting optimization problem is a nonconvex and nonlinear. The optimization problem is solved using the fmincon solver in Matlab and a custom made sequential optimization algorithm.
|Publication status||Published - 24 Sep 2017|
|Event|| 2017 ERNSI Workshop on System Identification - Domaine Lyon Saint Joseph, Lyon, France|
Duration: 24 Sep 2017 → 27 Sep 2017
|Workshop||2017 ERNSI Workshop on System Identification|
|Abbreviated title||ERNSI 2017|
|Period||24/09/17 → 27/09/17|