• Home
• What is IK?
• Focus Theme 2011
• Organization
• Program
  • IK 2011 Schedule
• Poster and Demo Session
• Location
• Stipends
• Registration
• Contact

Speaker:

Jörg Raisch

Disciplines:

control theory, systems theory, hybrid systems

Content:

I will start with briefly outlining the principles of feedback control and explain why feedback is ubiquitous both in nature and in technical applications. I will focus on synthesis (as opposed to analysis) of feedback control, i.e., on the question how a given dynamic behaviour can be changed in a desired manner via appropriately designed feedback.

I will then provide a brief survey on established control synthesis methods. Such methods are available for both continuous systems, i.e., systems evolving in Euclidean space, and for discrete event systems, i.e., systems evolving on finite (but unstructured) sets. We will then take a look at hybrid systems, which consist of nontrivially interacting continuous and discrete event components. It is a well known fact that, even for the most simple ingredients, discrete-continuous interaction can cause a surprising wealth of dynamic behaviour, ranging from chaotic to highly regular. Feedback design for hybrid systems is still a widely open area. I will focus on an abstraction-based approach, where the hybrid system to be controlled is first approximated by a discrete event system, for which we can subsequently use established design methods. Using an appropriate notion of abstraction, we will be able to formally guarantee that any feedback control system forcing the abstraction to satisfy a given set of specifications will do the same for the underlying hybrid system. This leaves us with the interesting question whether we can find a minimally complex abstraction that allows successful control synthesis for a given specification, or, in other words, what is the minimal required knowledge that is necessary to successfully control a system? To answer this question we will finally investigate a “learning by doing approach”. This is characterised by an iteration, where control synthesis, if it fails for an initially coarse abstraction, is followed by a diagnosis step that reveals the “knowledge bottlenecks” and allows for local, specification dependent refinement of the current abstraction.

CV:

Joerg Raisch studied  Engineering Cybernetics at  Stuttgart University
and Control Systems at  UMIST, Manchester, UK. After receiving his Ph.
D. from Stuttgart University, he spent two years as a postdoc at the
University of Toronto  and was subsequently awarded a  DFG (German
Science Foundation) fellowship to work towards his  "habilitation ".
From 2000 until 2006, he held the chair of "Systems Theory in
Engineering" at the  Otto-von-Guericke University Magdeburg. Since 2006,
he has held the chair of Control Systems at TU Berlin. He is also an
External Scientific Member of the Max Planck Institute for Dynamics of
Complex Technical Systems.