Awesome-LLM4AD
A curated list of awesome LLM/VLM/VLA/World Model for Autonomous Driving(LLM4AD) resources (continually updated)
What is Awesome-LLM4AD?
A curated list of awesome LLM/VLM/VLA/World Model for Autonomous Driving(LLM4AD) resources (continually updated)
Overview of Awesome-LLM4AD
What it is
This is a curated list of research papers and resources about Large Language Models for Autonomous Driving (LLM4AD). The repository is maintained by SJTU-ReThinklab and is continuously updated to track the frontier of LLM4AD, which encompasses Vision-Language Models for AD (VLM4AD) and Vision-Language-Action models for AD (VLA4AD) as integral components. The collection includes a survey paper available on arXiv (2311.01043) that provides more detailed discussions. The repository is licensed under Apache-2.0 and is open source.
Key features
- Curated collection of research papers on LLM, VLM, VLA, and World Models applied to autonomous driving.
- Papers are categorized by task: planning, perception, question answering, and generation.
- Each paper entry includes title, authors, publisher, task, keywords, code/project page, datasets, publish date, summary, and metrics.
- Includes a survey paper with detailed discussions and a citation format for referencing the repository.
- Repository is continuously updated to track the latest research.
Use cases
- Researchers and practitioners looking for a comprehensive overview of LLM applications in autonomous driving.
- Staying up-to-date with the latest papers and datasets in the field of LLM4AD, VLM4AD, and VLA4AD.
- Finding relevant papers for literature reviews, benchmarking, or project references.
- Understanding the motivation and taxonomy of LLM4AD, including the sim2real gap and long-tailed distribution challenges.
Getting started
See the repository README for instructions on how to navigate the curated list, contribute via pull requests, and cite the survey paper.
License and pricing
The repository is licensed under Apache-2.0 and is open source. Pricing is not stated.
Based on the project's public README and repository metadata.