14h00-15h00 – Salle du conseil (L2S)
Stability and stabilization of black-box switched linear systems: Probabilistic guarantees and sample complexity
Abstract. Data-driven methods have been quite a success in handling complex dynamical systems. In this presentation, I will talk about stability analysis of a particular class of hybrid systems, called switched linear systems, which consist of a finite set of linear dynamics (called modes) and a switching rule that indicates the current active mode of the system. Identification of such systems is often challenging due to hybrid behaviors, except for small systems with a small number of modes. To deal with this challenge, we propose data-driven methods in which we do not require a dynamical model. Instead, our methods only rely on a finite set of observations of system trajectories. To formally evaluate our methods, probabilistic guarantees are derived in the spirit of the scenario approach.
Biography. Zheming Wang is a postdoctoral research in the department of mathematical engineering at UCLouvain in Beligum. He is broadly interested in the area of control theory and optimization. His recent research interests focus on rigorous data-driven methods for analyzing and controlling complex systems with formal guarantees. He received a B.S. degree in Mechanical Engineering from Shanghai Jiao Tong University, China, in 2012, and a Ph.D. degree in Mechanical Engineering from National University of Singapore in 2016.
Let’s play! – Solving controller synthesis games for cyber-physical system design.
Abstract. Cyber-Physical Systems (CPS) are technical systems where a large software stack orchestrates the interaction of physical and digital components. Such systems are omnipresent in our daily life and their correct behavior is crucial. However, developing safe and reliable CPS is challenging. A promising direction towards this goal is the use of formal methods: automated methodologies that ensure system requirements during design-time. In my talk, I will discuss how infinite two-player games over finite graphs, originating from the formal methods community, can be utilized and enhanced for reliable CPS design. In particular, I will discuss how environment assumptions – used to model particularities of the system under control – can be efficiently incorporated into existing game-solving algorithms for controller synthesis.
Biography. Anne-Kathrin Schmuck studied Engineering Cybernetics at the OvGU Magdeburg in Germany from 2004 to 2009. During her studies she spent one year at UBC in Vancouver, Canada, working with Guy Dumont and six month at LTH in Lund, Sweden, working with Anders Rantzer. From 2010 to 2015 she was a research assistant in the Control Systems Group at TU Berlin working with Jörg Raisch. In March and December 2014 she was a visiting PhD student at the Cyber-Physical Systems Laboratory at UCLA working with Paulo Tabuada. She obtained my PhD degree from TU Berlin in September 2015. From 2015 to 2020 she was a postdoctoral researcher at MPI-SWS in Kaiserslautern working with Rupak Majumdar. In Fall 2020 she received an Emmy Noether Grant from the German Science Foundation (DFG) to establish her own research group at MPI-SWS to develop “Automated Modular Synthesis Techniques for Reliable Cyber Physical System Design”.