Master's Thesis: Investigating LLM- and RAG-Based Support for FMECA Creation in Robotic Systems
What to expect:
Nothing is perfect, not even a robotic system. Faults can happen in every of its components. Knowing a robotic system and being able to know where a fault comes from is a huge advantage and can increase the reliability of the robot. A common tool to collect the possible faults of a system is the “failure mode, effect and criticality analysis” (FMECA). Next to the faults, their cause, effect, detection and compensation provision is collected. Optionally, the severity and probability of the fault can be evaluated. Acquiring such an FMECA database can be a very laboursome manual work which draws a lot of expert resources. Therefore, it would be beneficial to automate the creation of the FMECA as far as possible. Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) can help to automize the workflow of system designers conducting the FMECA analysis of robotic systems, in particular in the space context. This thesis shall investigate whether, and to what extent, LLMs combined with RAG can meaningfully support or partially automate the creation of FMECA tables for robotic systems and which elements of the FMECA can be generated with sufficient reliability.
Your tasks will include:
- Understanding and analyzing the problem,
- Identifying constraints on inputs that might be used for RAG,
- Literature research on existing methods,
- Setting up the LLM and RAG pipeline,
- Evaluating the outcome with respect to practicability and reliability.
The successful candidate will have:
- an opportunity to learn from experienced roboticists in a professional environment,
- the potential to grow and apply your skills and knowledge.
What we expect from you:
- currently enrolled in a Master's degree in computer science, robotics, or a related field.
- Knowledge of fundamental concepts in LLMs and RAG.
- Advanced programming skills, basics of git.
- Professional working proficiency in English and/or German.
Further information:
- To apply for this position, please include a resume, a transcript of grades and a motivation letter with experience in relevant projects and/or classes.
- Earliest starting date: as soon as possible.
- Duration of contract: 6 months.
For application, first make contact (see below) to indicate your interest. We will then get back to you.
Kontakt
Office (AUF)