LLM-Empowered Event-Chain Driven Code Generation for ADAS in SDV systems

2025-11-26

AI Summary

This work addresses the challenge of automatically translating natural-language requirements into trustworthy automotive software code. We propose an event-chain-driven, retrieval-augmented code synthesis framework to simultaneously satisfy architectural correctness, behavioral consistency, and real-time feasibility in ADAS and software-defined vehicle (SDV) systems. Our method integrates large language models (LLMs) with retrieval-augmented generation (RAG) over vehicle signal specifications, explicitly modeling causal and temporal constraints. Code generation is guided by event-chain constraints, while signal mapping and formal verification mechanisms mitigate hallucination. Crucially, no LLM fine-tuning is required to ensure signal usage accuracy and system-level behavioral consistency. Evaluated on an emergency braking scenario, the framework generates executable, formally verified real-time code—demonstrating its effectiveness and engineering practicality for complex, safety-critical automotive applications.

Technology Category

• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Planning, Routing, and Scheduling: Planning with Language Models
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Search and Retrieval-Augmented AI: Large language models for search
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

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.

Problem

• Generates automotive code from natural language requirements
• Retrieves and validates vehicle signals to reduce LLM hallucinations
• Uses event chains to ensure behavioral consistency in code

Innovation

• RAG layer retrieves signals from VSS catalogs
• Event chains encode causal and timing constraints
• LLM-based code synthesis guided by event chains