Dear all,
Today's speaker at the Analysis and Probability seminar is Ludovic Sacchelli. The title of his talk is "Unobservability in system stabilization" and the abstract is below.
The talk will begin at 1:30 PM on Zoom, doors at 1:25 PM. The details are as follows:
Join Zoom Meeting
https://us02web.zoom.us/j/93945187422?pwd=MFJSZ0hLL014S1JvV2I1WjZ6Rm03UT09
Meeting ID: 939 4518 7422
Password: Monteith
You can keep up to date on possible future A/P talks here: https://calendar.uconn.edu/2020/month/04/376/.
Abstract: Stabilizing the state of a dynamical system to a target point is a classical problem in control theory. However, in many physical problems, only a partial measure of the state is known. A commonly used idea is to rely on an observer that dynamically learns the state of the system. To achieve stabilization via an observer, some guarantees on the quality of the measure are needed to make sure that the estimates are reliable.
A common hypothesis to achieve this strategy is "uniform observability" of the system, where for any state, a base level of information is always given by the measure. Nonlinear systems however, can present symmetries that completely break this assumption, and it is in fact not generic for a nonlinear dynamical system to be uniformly observable. Without a strong observability assumption, the usual strategies break down and new methods need to be explored to resolve this issue. Armed with case studies from quantum physics and system engineering, we develop a theory of embedding of systems. Considering high-dimensional (sometimes infinite-dimensional) embeddings of dynamical systems actually allows the introduction of new and better suited observability techniques.Best,
Sean