OpenDriveVLA: Towards End-to-end Autonomous Driving with Large Vision Language Action Model

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhou, Xingcheng, Han, Xuyuan, Yang, Feng, Ma, Yunpu, Tresp, Volker, Knoll, Alois
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918212876107776
author Zhou, Xingcheng
Han, Xuyuan
Yang, Feng
Ma, Yunpu
Tresp, Volker
Knoll, Alois
author_facet Zhou, Xingcheng
Han, Xuyuan
Yang, Feng
Ma, Yunpu
Tresp, Volker
Knoll, Alois
contents We present OpenDriveVLA, a Vision Language Action model designed for end-to-end autonomous driving, built upon open-source large language models. OpenDriveVLA generates spatially grounded driving actions by leveraging multimodal inputs, including 2D and 3D instance-aware visual representations, ego vehicle states, and language commands. To bridge the modality gap between driving visual representations and language embeddings, we introduce a hierarchical vision language alignment process, projecting both 2D and 3D structured visual tokens into a unified semantic space. Furthermore, we incorporate structured agent environment ego interaction modeling into the autoregressive decoding process, enabling the model to capture fine-grained spatial dependencies and behavior-aware dynamics critical for reliable trajectory planning. Extensive experiments on the nuScenes dataset demonstrate that OpenDriveVLA achieves state-of-the-art results across open-loop trajectory planning and driving-related question answering tasks. Qualitative analyses further illustrate its capability to follow high-level driving commands and generate trajectories under challenging scenarios, highlighting its potential for next-generation end-to-end autonomous driving.
format Preprint
id arxiv_https___arxiv_org_abs_2503_23463
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle OpenDriveVLA: Towards End-to-end Autonomous Driving with Large Vision Language Action Model
Zhou, Xingcheng
Han, Xuyuan
Yang, Feng
Ma, Yunpu
Tresp, Volker
Knoll, Alois
Computer Vision and Pattern Recognition
We present OpenDriveVLA, a Vision Language Action model designed for end-to-end autonomous driving, built upon open-source large language models. OpenDriveVLA generates spatially grounded driving actions by leveraging multimodal inputs, including 2D and 3D instance-aware visual representations, ego vehicle states, and language commands. To bridge the modality gap between driving visual representations and language embeddings, we introduce a hierarchical vision language alignment process, projecting both 2D and 3D structured visual tokens into a unified semantic space. Furthermore, we incorporate structured agent environment ego interaction modeling into the autoregressive decoding process, enabling the model to capture fine-grained spatial dependencies and behavior-aware dynamics critical for reliable trajectory planning. Extensive experiments on the nuScenes dataset demonstrate that OpenDriveVLA achieves state-of-the-art results across open-loop trajectory planning and driving-related question answering tasks. Qualitative analyses further illustrate its capability to follow high-level driving commands and generate trajectories under challenging scenarios, highlighting its potential for next-generation end-to-end autonomous driving.
title OpenDriveVLA: Towards End-to-end Autonomous Driving with Large Vision Language Action Model
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2503.23463