Configuring the Numerical Optimization of the D-optimal Input Design problem for Block Structured Systems

Research output: Unpublished contribution to conferencePoster

Abstract

The field of optimal input design considers the problem of finding the most informative input signal out of the set of possible excitation signals given some prior knowledge about the system. In its most general form, finding an optimal input signal comes down to solving an optimization problem in which a scalar measure of the Fisher information matrix is maximized with respect to the parametrization of the input sequence. The complexity of this optimization problem strongly depends on the model structure, the input parametrization and the properties of the scalar information measure.
In this study, the optimal input design is computed for a block structured systems, which consist of a combination of linear time invariant blocks and static nonlinear blocks. The class of inputs is restricted to periodic bandlimited excitations and is parameterized in the time domain. Under these assumptions, the resulting optimization problem becomes nonlinear and non-convex. This type of optimization is very sensitive with respect to the solver and the design choices. Based on extensive simulation results, guidelines are extracted for the different design choices that have to be made during the optimization.
Original languageEnglish
Publication statusPublished - 25 Sep 2016
EventERNSI WORKSHOP 2016 - Cison di Valmarino, Italy
Duration: 25 Sep 201628 Sep 2016

Workshop

WorkshopERNSI WORKSHOP 2016
CountryItaly
CityCison di Valmarino
Period25/09/1628/09/16

Fingerprint

Dive into the research topics of 'Configuring the Numerical Optimization of the D-optimal Input Design problem for Block Structured Systems'. Together they form a unique fingerprint.

Cite this