Einladung zum Vortrag im Kolloquium Technische Kybernetik

Optimal control and particle filtering
for nonlinear descriptor systems

Johan Sjöberg, M.Sc.
Department of Electrical Engineering
Linköping University, Linköping, Sweden

    Zeit: Dienstag · 7. 11. 2006 · 16:00 Uhr
    Ort: IST-Seminarraum 3.241 · Pfaffenwaldring 9 · Campus Stuttgart-Vaihingen

Abstract

In this talk, optimal feedback control and particle filtering for nonlinear descriptor systems will be discussed. The main motivation for studying this class of models is the growing use of object-oriented modeling tools, often yielding descriptions in this form.

Optimal feedback control is normally computed by solving the Hamilton-Jacobi-Bellman equation. A problem when computing the optimal feedback law using this equation is that it involves solving a nonlinear partial differential equation, which often cannot be solved explicitly. However, Lukes etc. showed that for nonlinear state-space systems, it can still be possible to find a local optimal solution, expressed as power series expansions. We will show that this method can be extended to handle also rather general descriptor systems.

Particle filtering for estimating the state of nonlinear state-space systems has got a lot of attention recently. This talk will be concluded with a discussion about how particle filtering can be extended to also handle nonlinear descriptor systems.

Biographical Information