AI-based extraction of requirements from regulations for automotive engineering
DS 125: Proceedings of the 34th Symposium Design for X (DFX2023)
Year: 2023
Editor: Dieter Krause, Kristin Paetzold-Byhain, Sandro Wartzack
Author: Iris Gräßler, Deniz Ozcan, Daniel Preuß
Series: DfX
Institution: Heinz Nixdorf Institute, Chair of Product Creation, Paderborn University
Page(s): 163-172
DOI number: 10.35199/dfx2023.17
Abstract
Automotive engineering requires compliance with regulations for certification. In specifications, regulations are referenced, which need to be analyzed manually to elicit requirements. This process is time-consuming and leads to high costs. The aim of this research is to evaluate artificial intelligence (AI) models in terms of extracting requirements automatically from regulations. Relevant AI models are identified in a systematic literature analysis and evaluated using success criteria. The most promising AI models are implemented in a pipeline for requirements extraction. The performance of these AI models is assessed in a comparative study using automotive regulations. The results show which AI models are best suited for this task.
Keywords: Requirements Engineering, Natural Language Processing, Artificial Intelligence, Automotive Engineering