Publication
We position OpenDriveLab as one of the most top research teams around globe, since we've got talented people and published work at top venues.
Li Chen,
Chonghao Sima,
Antonio Loquercio,
Preprint 2026Position Paper
Modi Shi,
Li Chen,
Chiming Liu,
Guanghui Ren,
Di Huang,
Maoqing Yao,
arXiv 2025
The first comprehensive analysis of data diversity principles revealing optimal scaling strategies for large-scale robotic manipulation training.
Guanghui Ren,
Maoqing Yao,
RSS 2025
A unified vision-language-action framework that enables policy learning across different environments.
Yihan Hu,
Li Chen,
Keyu Li,
Chonghao Sima,
Xizhou Zhu,
Siqi Chai,
Senyao Du,
Tianwei Lin,
Lewei Lu,
Qiang Liu,
Jifeng Dai,
Team AgiBot-World
IROS 2025 Best Paper Finalist
A novel generalist policy that leverages latent action representations to maximize data utilization, demonstrating predictable performance scaling with increased data volume.
Li Chen,
Bernhard Jaeger,
IEEE-TPAMI 2024Survey
In this survey, we provide a comprehensive analysis of more than 270 papers on the motivation, roadmap, methodology, challenges, and future trends in end-to-end autonomous driving.
Li Chen,
Modi Shi,
Yanjie Zhang,
Delong Li,
Chuanzhe Suo,
Chuang Wang,
arXiv 2025
A unified VLA framework enabling large-space humanoid loco-manipulation via unified latent learning and loco-manipulation-oriented RL. 
Yixuan Pan,
Ruoyi Qiao,
Li Chen,
Liang Pan,
Haoguang Mai,
Hao Zhao,
Cunyuan Zheng,
arXiv 2025
A unified whole-body control policy for humanoid robots that enables zero-shot execution of diverse motions, including Ip Man'squat, dancing, running and real-time teleoperation.
Li Chen,
Jia Zeng,
Heming Cui,
Maoqing Yao,
arXiv 2024
Our objective is to develop a synergistic dual-system framework which supplements the generalizability of large-scale pre-trained generalist with the efficient and task-specific adaptation of specialist.
Jia Zeng,
Li Chen,
Yanchao Yang,
Guyue Zhou,
Heming Cui,
NeurIPS 2024
CLOVER employs a text-conditioned video diffusion model for generating visual plans as reference inputs, then these sub-goals guide the feedback-driven policy to generate actions with an error measurement strategy.
Longyan Wu,
Jieji Ren,
Li Chen,
Ran Huang,
Guoying Gu,
arXiv 2025
A human-centric and robot-free visuo-tactile data collection system for high-quality and efficient robot manipulation.
Jia Zeng,
Wenke Xia,
Li Chen,
Hao Dong,
Haoming Song,
Dong Wang,
Di Hu,
Heming Cui,
Bin Zhao,
Xuelong Li,
RSS 2024
We propose a general pre-training pipeline that learns Manipulation by Predicting the Interaction (MPI).
Guang Li,
Junli Wang,
Yinfeng Gao,
Zhang Zhang,
Liang Wang,
Hangjun Ye,
Tieniu Tan,
Long Chen,
arXiv 2025
A scalable sim-real learning framework that synthesizes high-fidelity driving data and cboosts end-to-end planners to achieve robust, generalizable autonomy with principled scaling insights.
Haohan Yang,
Li Chen,
Ke Guo,
Hongchen Li,
Chen Lv
IEEE-TPAMI 2026
Chonghao Sima,
Katrin Renz,
Li Chen,
ECCV 2024 Oral
Unlocking the future where autonomous driving meets the unlimited potential of language.
Long Chen,
Yuqian Shao,
Xiangyu Yue,
Li Chen
NeurIPS 2025 Spotlight
ReSim is a driving world model that enables Reliable Simulation of diverse open-world driving scenarios under various actions, including hazardous non-expert ones. A Video2Reward model estimates the reward from ReSim's simulated future.
William Ljungbergh,
Hongzi Zhu,
Christoffer Petersson,
ICCV 2025
Chonghao Sima,
Zhiding Yu,
Shiyi Lan,
Jose M. Alvarez
arXiv 2025
Li Chen,
Jun Zhang,
NeurIPS 2024
A generalizable driving world model with high-fidelity open-world prediction, continuous long-horizon rollout, and zero-shot action controllability.
Li Chen,
Chen Lv,
NeurIPS 2024
Daniel Dauner,
Marcel Hallgarten,
Xinshuo Weng,
Zhiyu Huang,
Igor Gilitschenski,
Boris Ivanovic,
Marco Pavone,
NeurIPS 2024 Track Datasets and Benchmarks
Data-Driven Non-Reactive Autonomous Vehicle Simulation and Benchmarking.
Li Chen,
Bo Dai,
Jia Zeng,
Jun Zhang,
CVPR 2024 Highlight
We aim to establish a generalized video prediction paradigm for autonomous driving by presenting the largest multimodal driving video dataset to date, OpenDV-2K, and a generative model that predicts the future given past visual and textual input, GenAD.
Li Chen,
Yanan Sun,
CVPR 2024 Highlight
A new self-supervised pre-training task for end-to-end autonomous driving, predicting future point clouds from historical visual inputs, joint modeling the 3D geometry and temporal dynamics for simultaneous perception, prediction, and planning.
Yang Li,
Jia Zeng,
Huilin Xu,
Pinlong Cai,
Li Chen,
Feng Xu,
Lu Xiong,
Jingdong Wang,
Futang Zhu,
Kai Yan,
Chunjing Xu,
Tiancai Wang,
Fei Xia,
Beipeng Mu,
Dahua Lin,
中國科學：信息科學Survey
Jia Zeng,
Hang Qiu,
Hongzi Zhu,
Minyi Guo,
ECCV 2024
Revive driving scene understanding by delving into the embodiment philosophy.
Li Chen,
Patrick Langechuan Liu,
ICCV 2023 Oral
A new paradigm for end-to-end autonomous driving without causal confusion issue.
Li Chen,
Jiangwei Xie,
Conghui He,
CVPR 2023
A scalable decoder paradigm that generates the future trajectory and action of the ego vehicle for end-to-end autonomous driving.
Zhenjie Yang,
arXiv 2023Survey
A collection of research papers about LLM-for-Autonomous-Driving (LLM4AD).
Li Chen,
ICLR 2023
An intuitive and straightforward fully self-supervised framework curated for the policy pre-training in visuomotor driving.
Trajectory-guided Control Prediction for End-to-end Autonomous Driving: A Simple yet Strong Baseline
Li Chen,
NeurIPS 2022
Take the initiative to explore the combination of controller based on a planned trajectory and perform control prediction.
Li Chen,
Dacheng Tao
ECCV 2022
A spatial-temporal feature learning scheme towards a set of more representative features for perception, prediction and planning tasks simultaneously.
Shihao Li,
Tuo An,
Peng Su,
Boyang Wang,
Haiou Liu,
Chen Lv,
arXiv 2025
OMEGA is a training-free, optimization-guided diffusion framework that enforces structural consistency and interaction reasoning during sampling to generate realistic, controllable, and safety-critical multi-agent driving scenes.
Zhenhua Wu,
Carl Lindström,
Peng Su,
Matthias Nießner,
arXiv 2025
Peijin Jia,
Li Chen,
Kun Jiang,
ICLR 2024
We advocate Lane Segment as a map learning paradigm that seamlessly incorporates both map geometry and topology information.
Haisong Liu,
Yang Chen,
Haiguang Wang,
Jia Zeng,
Li Chen,
Limin Wang
ECCV 2024
Chonghao Sima,
Jingdong Wang,
NeurIPS 2023
The unified framework to enhance 3D object detection by uniting a multi-modal expert model with a trajectory distillation module.
Yang Li,
Li Chen,
Chonghao Sima,
Zhenbo Liu,
Peijin Jia,
Yuting Wang,
Shengyin Jiang,
Feng Wen,
Hang Xu,
Wei Zhang,
NeurIPS 2023 Track Datasets and Benchmarks
The world's first perception and reasoning benchmark for scene structure in autonomous driving.
Chonghao Sima,
Jifeng Dai,
Lewei Lu,
Jia Zeng,
Hanming Deng,
Hao Tian,
Enze Xie,
Jiangwei Xie,
Li Chen,
Yang Li,
Si Liu,
Jianping Shi,
Dahua Lin,
IEEE-TPAMI 2023Survey
We review the most recent work on BEV perception and provide analysis of different solutions.
Chonghao Sima,
Wenwen Tong,
Tai Wang,
Li Chen,
Silei Wu,
Hanming Deng,
Yi Gu,
Lewei Lu,
Dahua Lin,
ICCV 2023
Occupancy serves as a general representation of the scene and could facilitate perception and planning in the full-stack of autonomous driving.
Chonghao Sima,
Shaoshuai Shi,
Shangzhe Di,
Si Liu,
IROS 2023
We propose Sparse Dense Fusion (SDF), a complementary framework that incorporates both sparse-fusion and dense-fusion modules via the Transformer architecture.
Li Chen,
Yu Liu,
IEEE-TPAMI 2023
HDGT formulates the driving scene as a heterogeneous graph with different types of nodes and edges.
Jia Zeng,
Li Chen,
Hanming Deng,
Lewei Lu,
CVPR 2023
We investigate on how to distill the knowledge from an imperfect expert. We propose FD3D, a Focal Distiller for 3D object detection.
Chenyu Yang,
Yuntao Chen,
Hao Tian,
Chenxin Tao,
Xizhou Zhu,
Zhaoxiang Zhang,
Gao Huang,
Lewei Lu,
Jie Zhou,
Jifeng Dai
CVPR 2023 Highlight
A novel bird's-eye-view (BEV) detector with perspective supervision, which converges faster and better suits modern image backbones.
Li Chen,
Yang Li,
Shengyin Jiang,
Yuting Wang,
Hang Xu,
Chunjing Xu,
arXiv 2023
A new baseline for scene topology reasoning, which unifies heterogeneous feature learning and enhances feature interactions via the graph neural network architecture and the knowledge graph design.
Jia Zeng,
Shengchuan Zhang,
Liujuan Cao,
Rongrong Ji,
arXiv 2023
We propose GAPretrain, a plug-and-play framework that boosts 3D detection by pretraining with spatial-structural cues and BEV representation.
Wenwen Tong,
Jiangwei Xie,
Yang Li,
Hanming Deng,
Bo Dai,
Lewei Lu,
Hao Zhao,
PRCV 2024
We propose a 3D data augmentation approach termed Drive-3DAug to augment the driving scenes on camera in the 3D space.
Li Chen,
Jia Zeng,
CoRL 2022
We find taking scratch trajectories generated by MLP as input, a refinement module based on structures with temporal prior, could boost the accuracy.
BEVFormer: Learning Bird's-Eye-View Representation From LiDAR-Camera via Spatiotemporal Transformers
Enze Xie,
Chonghao Sima,
Tong Lu,
Jifeng Dai
IEEE-TPAMI 2025
A paradigm for autonomous driving that applies both Transformer and Temporal structure to generate BEV features.
Li Chen,
Chonghao Sima,
Yang Li,
Zehan Zheng,
Conghui He,
Jianping Shi,
ECCV 2022 Oral
PersFormer adopts a unified 2D/3D anchor design and an auxiliary task to detect 2D/3D lanes; we release one of the first large-scale real-world 3D lane datasets, OpenLane.
Qingsong Yao,
Yanan Sun,
Renrui Zhang,
Hao Zhao,
Hongzi Zhu,
ICCV 2025
DetAny3D is a promptable 3D detection foundation model that leverages 2D foundation knowledge to enable zero-shot 3D object detection from monocular images across novel categories and camera settings.
Yanan Sun,
Li Chen,
Fei Xia,
IEEE-TPAMI 2026
Yifei Zhang,
Hao Zhao,
Siheng Chen
CVPR 2024
Shilin Yan,
Renrui Zhang,
Ziyu Guo,
Wenchao Chen,
Wei Zhang,
Zhongjiang He,
Peng Gao
AAAI 2024
Peng Gao,
Hao Sun,
Houqiang Li,
Jiebo Luo
CVPR 2023
An efficient MIM paradigm MaskAlign and a Dynamic Alignment module to apply learnable alignment to tackle the problem of input inconsistency.
Peng Gao,
Renrui Zhang,
Rongyao Fang,
Ziyi Lin,
Hongsheng Li,
Qiao Yu
IJCV 2023
Introducing high-level and low-level representations to MAE without interference during pre-training.
Shaofeng Zhang,
Lyn Qiu,
Feng Zhu,
Hengrui Zhang,
Rui Zhao,
Xiaokang Yang
CVPR 2022