LLM-Empowered Event-Chain Driven Code Generation for ADAS in SDV systems
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911289750585344 |
|---|---|
| author | Petrovic, Nenad Kroth, Norbert Torschmied, Axel Song, Yinglei Pan, Fengjunjie Zolfaghari, Vahid Purschke, Nils Kirchner, Sven Wu, Chengdong Schamschurko, Andre Zhang, Yi Knoll, Alois |
| author_facet | Petrovic, Nenad Kroth, Norbert Torschmied, Axel Song, Yinglei Pan, Fengjunjie Zolfaghari, Vahid Purschke, Nils Kirchner, Sven Wu, Chengdong Schamschurko, Andre Zhang, Yi Knoll, Alois |
| contents | This paper presents an event-chain-driven, LLM-empowered workflow for generating validated, automotive code from natural-language requirements. A Retrieval-Augmented Generation (RAG) layer retrieves relevant signals from large and evolving Vehicle Signal Specification (VSS) catalogs as code generation prompt context, reducing hallucinations and ensuring architectural correctness. Retrieved signals are mapped and validated before being transformed into event chains that encode causal and timing constraints. These event chains guide and constrain LLM-based code synthesis, ensuring behavioral consistency and real-time feasibility. Based on our initial findings from the emergency braking case study, with the proposed approach, we managed to achieve valid signal usage and consistent code generation without LLM retraining. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_21877 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | LLM-Empowered Event-Chain Driven Code Generation for ADAS in SDV systems Petrovic, Nenad Kroth, Norbert Torschmied, Axel Song, Yinglei Pan, Fengjunjie Zolfaghari, Vahid Purschke, Nils Kirchner, Sven Wu, Chengdong Schamschurko, Andre Zhang, Yi Knoll, Alois Software Engineering Artificial Intelligence This paper presents an event-chain-driven, LLM-empowered workflow for generating validated, automotive code from natural-language requirements. A Retrieval-Augmented Generation (RAG) layer retrieves relevant signals from large and evolving Vehicle Signal Specification (VSS) catalogs as code generation prompt context, reducing hallucinations and ensuring architectural correctness. Retrieved signals are mapped and validated before being transformed into event chains that encode causal and timing constraints. These event chains guide and constrain LLM-based code synthesis, ensuring behavioral consistency and real-time feasibility. Based on our initial findings from the emergency braking case study, with the proposed approach, we managed to achieve valid signal usage and consistent code generation without LLM retraining. |
| title | LLM-Empowered Event-Chain Driven Code Generation for ADAS in SDV systems |
| topic | Software Engineering Artificial Intelligence |
| url | https://arxiv.org/abs/2511.21877 |