Master Thesis or Working Student (f/m/x) Development AI Fault Detection Framework Train Localization

What to expect:

At the DLR Institute of Communications and Navigation we conduct research on different methods for precise and reliable localization of trains. To evaluate these methods, we have collected data from different kinds of sources (e. g. video cameras, GNSS receivers, magnetic field sensors, accelerometers). To evaluate our data in an efficient way, we want to build a framework to compare data from different sources with a focus on detecting and explaining sensor faults.

Your tasks:

Depending on your qualifications, your tasks can include:

• Programming a unified framework containing sensor and video data from different sources
• Integration of AI tools (e. g. video object detectors like YOLO or open-source LLMs) into the toolchain and/or adaption of AI tools to the railway domain
• Evaluation of measurement faults in sensor data using the developed framework
• Application of fault detection methods to train localization

Your profile:

• Study of Computer Science, Software Engineering, Information Technology, Data Science, Geodesy or similar
• good programming skills in Python and/or Matlab
• Hands-on experience with machine learning tools
• Knowledge about fault detection methods is not necessary, but can be beneficial
• Ability to work independently