Towards Automated Safety Requirements Derivation Using Agent-based RAG
Balahari Vignesh Balu, Florian Geissler, Francesco Carella, Joao-Vitor Zacchi · 2025
Read Towards Automated Safety Requirements Derivation Using Agent-based RAG by Balahari Vignesh Balu, Florian Geissler, Francesco Carella, Joao-Vitor Zacchi (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-11243.txt.
Description
Balahari Balu and co-authors automate the writing of safety requirements for self-driving cars, in a 2025 paper. Safety analysis normally depends on experts grinding through standards by hand. Language models can draft requirements, but off-the-shelf models lack domain knowledge. Retrieval systems fix that gap, yet they falter on complex queries where the right passage is hard to fetch. The authors add an agent layer that plans and refines the search.
The system was built on a pool of automotive standards documents and tested on a public automated-driving case study. Retrieved material matched the queries better than with plain retrieval. In safety work, a missed clause is not a style issue, so relevance gains translate directly into more trustworthy drafts for engineers to review.
Catalog record
| CREATOR | Balahari Vignesh Balu, Florian Geissler, Francesco Carella, Joao-Vitor Zacchi |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 31 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-11243