Prof. Martin Fränzle
Department of Computing Science
Carl von Ossietzky University of Oldenburg, Germany
Tuesday 2019-01-29 16:00
IST-Seminar-Room 2.255 - Pfaffenwaldring 9 - Campus Stuttgart-Vaihingen
The advent of systems of cooperative cyber-physical systems draws attention to a central problem of networked and distributed control systems: the ubiquity of delay in feedback loops between logically or spatially distributed components, which is not adequately reflected in traditional models of hybrid-state dynamics based on ordinary differential equations and immediate transitions. Occurrence of feedback delays may significantly alter a system's dynamic response. Unmodeled delays in a control loop consequently have the potential to invalidate any stability and safety certificate obtained on a related delay-free model, which is current practice in hybrid-system analysis. In this talk, we will present various approaches to the analysis and correct-by-construction design of dynamical systems subject to delayed information exchange, as pertinent to distributed hybrid systems. We will explain automatic verification procedures for invariance properties and bounded temporal-logic based on constraint-solving or rigorous generalization from simulations. This analytical view will be complemented by a constructive one based on a notion of delayed games and corresponding strategy synthesis algorithms.
Martin Fränzle is Professor for Hybrid Discrete-Continuous Systems at the Carl von Ossietzky Universität Oldenburg, Germany. Previous positions include an associate professorship as a Velux visiting professorship at the Technical University of Denmark, post-doctoral position at Oldenburg and Kiel, an a doctoral researcher position at Kiel, where he also obtained his doctoral degree. His research interests are in modelling, verification, and synthesis of reactive, real-time, and hybrid dynamics in embedded and cyber-physical systems. He has worked on the semantics of high-level modelling and specification languages and on decision problems and their application to verifying and synthesizing real-time and hybrid discrete-continuous systems including settings subject to stochastic disturbances.