Realistic Corner Case Generation for Autonomous Vehicles with Multimodal Large Language Model
Qiujing Lu, Meng Ma, Ximiao Dai, Xuanhan Wang, Shuo Feng
TL;DR
AutoScenario tackles the challenge of generating realistic, diverse, and controllable corner cases for autonomous vehicles by leveraging a multimodal LLM-powered pipeline that translates multimodal real-world data into textual scene descriptions and drives SUMO and CARLA to construct simulation-ready scenarios. The approach encodes inputs into a unified linguistic space and decodes them into network, agent, and object configurations guided by domain knowledge, enabling tailored testing scenarios from text, images, and videos, including crash reports. Experiments show high conformity, broad diversity, and the ability to produce novel corner cases, with ablations confirming the necessity of interpreters, chain-of-thought prompting, and prior knowledge. The work advances scenario-based AV testing by bridging narrative requirements and detailed simulation configurations, and suggests future work on photorealism and online learning to further adapt to corner-case discovery.
Abstract
To guarantee the safety and reliability of autonomous vehicle (AV) systems, corner cases play a crucial role in exploring the system's behavior under rare and challenging conditions within simulation environments. However, current approaches often fall short in meeting diverse testing needs and struggle to generalize to novel, high-risk scenarios that closely mirror real-world conditions. To tackle this challenge, we present AutoScenario, a multimodal Large Language Model (LLM)-based framework for realistic corner case generation. It converts safety-critical real-world data from multiple sources into textual representations, enabling the generalization of key risk factors while leveraging the extensive world knowledge and advanced reasoning capabilities of LLMs.Furthermore, it integrates tools from the Simulation of Urban Mobility (SUMO) and CARLA simulators to simplify and execute the code generated by LLMs. Our experiments demonstrate that AutoScenario can generate realistic and challenging test scenarios, precisely tailored to specific testing requirements or textual descriptions. Additionally, we validated its ability to produce diverse and novel scenarios derived from multimodal real-world data involving risky situations, harnessing the powerful generalization capabilities of LLMs to effectively simulate a wide range of corner cases.