Ongoing Project

DFG Projekt "SApPHIRe: A Systematic Approach for Provably correct Hybrid systems that are Intelligent and Resilient"

Cyber-physical systems increasingly run safety-critical applications, from energy supply to self-driving vehicles. Verifying them is hard: they combine discrete and continuous behavior, face uncertainty, and increasingly rely on learning. Deductive methods offer rigorous, modular guarantees but often neglect uncertainty, while quantitative methods handle uncertainty but are limited to specific system classes and time-bounded properties.

SapPHIRe bridges this gap with compositional, quantitative verification of hybrid systems under uncertainty. It develops new modeling techniques for uncertainty, decomposes system-level requirements into context-sensitive service levels, and safely integrates learning components.

DFG Project "RealySt: Reachability Analysis for Stochastic Hybrid Systems" in cooperation with Prof. Dr. Erika Ábrahám from RWTH Aachen University

 

NWO Projekt in Kooperation mit Johann Hurink, University of Twente:

NWO - ISOLATE

Integrated distributed SCADA security through local approximations of power flow equations

On april 1, 2021 Verena Menzel starts working on the project as a doctoral student.

‘Cybersecurity in future energy distribution grids’.

Renewable energy requires advanced control networks. To ensure safe and secure energy distribution in complex energy grids, we build and implement an additional security layer: Using the grid topology and real-time measurements from neighbouring stations, a model of the state of substations is maintained, which allows to detect malicious commands and false measurements.

 

Finished Projects

NWO Veni: Counting on a reliable water supply (laymen title).

ALLEGIO: Composable Embedded Systems for Healthcar.

MOSES (NWO): More secure SCADA networks through self-awareness.

ROCKS: RigorOus dependability analysis using model ChecKing techniques for Stochastic systems.

MatMam: Mean-Filed Approximation techniques for Markov Models.