LMDrive: Closed-Loop End-to-End Driving with Large Language Models

Hao Shao^1,2  Yuxuan Hu³  Letian Wang⁴

Steven L. Waslander⁴  Yu Liu^2,5  Hongsheng Li^1,3,5

¹CUHK MMLab  ²SenseTime Research  ³CPII under InnoHK

⁴University of Toronto  ⁵Shanghai Artificial Intelligence Laboratory

Abstract

Despite significant recent progress in the field of autonomous driving, modern methods still struggle and can incur serious accidents when encountering long-tail un foreseen events and challenging urban scenarios. On the one hand, large language models (LLM) have shown impressive reasoning capabilities that approach “Artificial General Intelligence”. On the other hand, previous autonomous driving methods tend to rely on limited-format inputs (e.g. sensor data and navigation waypoints), restricting the vehicle’s ability to understand language information and interact with humans. To this end, this paper introduces LMDrive, a novel language-guided, endto-end, closed-loop autonomous driving framework. LM Drive uniquely processes and integrates multi-modal sensor data with natural language instructions, enabling interaction with humans and navigation software in realistic instructional settings. To facilitate further research in language-based closed-loop autonomous driving, we also publicly release the corresponding dataset which includes approximately 64K instruction-following data clips, and the LangAuto benchmark that tests the system’s ability to handle complex instructions and challenging driving scenarios. Extensive closed-loop experiments are conducted to demonstrate LMDrive’s effectiveness. To the best of our knowledge, we’re the very first work to leverage LLMs for closed-loop end-to-end autonomous driving. Codes can be found at our webpage.