Text2Scenario: Text-Driven Scenario Generation for Autonomous Driving Test
🤖 AI Summary
To address the low efficiency and poor scalability of manually constructing segmented scenarios in autonomous driving (AD) simulation testing, this paper proposes a text-driven end-to-end scenario generation framework. Our method introduces hierarchical prompt engineering—synergizing large language models (LLMs) with domain-specific languages (DSLs)—to automatically parse natural-language scenario descriptions and serialize them into executable test scenarios. It further integrates a hierarchical scenario knowledge base with a semantic matching mechanism to ensure structural validity and intent consistency. Experimental results demonstrate that the generated scenarios achieve over 92% intent compliance rate, significantly improving scenario construction efficiency. The framework enables precise, large-scale, automated evaluation across diverse AD software stacks, supporting scalable and intent-preserving scenario synthesis for robust AD validation.
Application Category
📝 Abstract
Autonomous driving (AD) testing constitutes a critical methodology for assessing performance benchmarks prior to product deployment. The creation of segmented scenarios within a simulated environment is acknowledged as a robust and effective strategy; however, the process of tailoring these scenarios often necessitates laborious and time-consuming manual efforts, thereby hindering the development and implementation of AD technologies. In response to this challenge, we introduce Text2Scenario, a framework that leverages a Large Language Model (LLM) to autonomously generate simulation test scenarios that closely align with user specifications, derived from their natural language inputs. Specifically, an LLM, equipped with a meticulously engineered input prompt scheme functions as a text parser for test scenario descriptions, extracting from a hierarchically organized scenario repository the components that most accurately reflect the user's preferences. Subsequently, by exploiting the precedence of scenario components, the process involves sequentially matching and linking scenario representations within a Domain Specific Language corpus, ultimately fabricating executable test scenarios. The experimental results demonstrate that such prompt engineering can meticulously extract the nuanced details of scenario elements embedded within various descriptive formats, with the majority of generated scenarios aligning closely with the user's initial expectations, allowing for the efficient and precise evaluation of diverse AD stacks void of the labor-intensive need for manual scenario configuration. Project page: https://caixxuan.github.io/Text2Scenario.GitHub.io.
Problem
Research questions and friction points this paper is trying to address.
Automates generation of autonomous driving test scenarios
Reduces manual effort in scenario configuration
Enhances precision in evaluating AD technologies
Innovation
Methods, ideas, or system contributions that make the work stand out.
LLM-based scenario generation from text
Hierarchical scenario repository integration
Domain Specific Language for executable scenarios
🔎 Similar Papers
No similar papers found.
