About Me

Hi, I am Michaela, a research engineer at Bosch Corporate Research (Robert Bosch GmbH) in Renningen, close to Stuttgart. I joined Bosch Research in January 2023. My focus lies in verification of planning and decision making in autonomous systems. I am working on model-checking robotic deliberation and autonomous driving behavior. Therefore, I am associated with the robotics and autonomous driving portfolios at Bosch Research. From having worked with different model types and state-of-the-art planning and verification tools, I have a broad and solid knowledge of model-driven engineering methods. Given my passion for cutting-edge research, I love to advance innovative verification methods for automated planning and decision making of highly autonomous systems.

Since January 2024, I am the activity lead of an internal project in the robotics portfolio, where I am responsible for the project and personnel management, as well as for the technical focus and implementation of the project goals. In addition, I am a PI of the EU Horizon project CONVINCE, where our focus lies on model-checking robotic deliberation to achieve robust robot behavior using statistical model checking, situation understanding, and learning (used tooling & formats: JANI, Storm, Prism, temporal logics, ROS 2). I am supervising PhD students, e.g., working on formally verifying behavior trees, Master’s thesis students, etc. Furthermore, I am quite active in the internal and external presentation of and communication for the project.

In the autonomous driving portfolio I was working on formally verifying a behavior planner of an autonomous car with the help of model checking in nuXmv. Learn more about this project in the Bosch Research Blog. The corresponding TACAS paper was honored with the Distinguished Artifact Award.

My intrinsic motivation always pushed new ideas forward and my proactive working style opened doors for fruitful collaborations with other scientists all over the world. I deeply enjoy working together in strong teams with complementary skills. Many of my papers have been developed in teams of researchers from different backgrounds.

From November 2017 to December 2022, I have been a research associate at the chair of Prof. Dr.-Ing. Holger Hermanns at the Dependable Systems and Software Group at Saarland University. There, I have been working in the area of quantitative verification and Markov decision processes, especially at the interface of these topics to automatic planning and the AI community. In June 2022, I successfully defended my PhD (Dr.-Ing.) thesis entitled “On the Connection of Probabilistic Model Checking, Planning, and Learning for System Verification”. The main contributions of my thesis are (1) a novel quantitative verification approach, called Modysh, based on automated probabilistic planning methods implemented in a state-of-the-art model checker, (2) a lightweight verification technique for neural network decision making agents, called Deep Statistical Model Checking (DSMC), mainly evaluated on an abstract autonomous driving benchmark designed by me, as well as (3) the design of models and benchmarks usable for evaluation and testing purposes in the model checking, verification, and planning communities.

I worked with state-of-the-art probabilistic model checking and planning tools & formats (The Modest Toolset, Prism, Storm, JANI, FastDownward, (P)PDDL), designed properties in temporal logics, and developed new verification techniques for Markov decision processes (MDPs) of cyber-physical systems. In addition, I was in charge of the implementation of algorithms in state-of-the-art probabilistic model checking tools (The Modest Toolset, Momba).

I studied Computer Science at Saarland University where I received my Master’s degree (honor’s degree, Günter-Hotz-Medal) in March 2018 and my Bachelor’s degree in July 2016.

If you are interested in my work, explore the other tabs or feel free to contact me.