OpenDriveLab

Planning-oriented Autonomous Driving

Yihan Hu, Jiazhi Yang, Li Chen, Keyu Li, Chonghao Sima, Hongyang Li, et al.
CVPR 2023 Award Candidate [Sell a New Philosophy]
A comprehensive framework up-to-date that incorporates full-stack driving tasks in one network.

BEVFormer: Learning Bird's-Eye-View Representation from Multi-Camera Images via Spatiotemporal Transformers

Zhiqi Li, Wenhai Wang, Hongyang Li, Enze Xie, Chonghao Sima, et al.
ECCV 2022 [Baseline][nuScenes First Place]
[Waymo Challenge 2022 Official First Place]
A paradigm for autonomous driving that applies both Transformer and Temporal structure to generate BEV features.

PersFormer: 3D Lane Detection via Perspective Transformer and the OpenLane Benchmark

Li Chen, Chonghao Sima, Yang Li, Xiangwei Geng, Junchi Yan, et al.
ECCV 2022 (Oral) [Redefine the Community]
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.

Policy Pre-Training for End-to-End Autonomous Driving via Self-Supervised Geometric Modeling

Penghao Wu, Li Chen, Hongyang Li, Xiaosong Jia, Junchi Yan, Yu Qiao
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

Penghao Wu, Xiaosong Jia, Li Chen, Junchi Yan, Hongyang Li, Yu Qiao
NeurIPS 2022 [Carla First Place]
Take the initiative to explore the combination of controller based on a planned trajectory and perform control prediction.

ST-P3: End-to-end Vision-based Autonomous Driving via Spatial-Temporal Feature Learning

Shengchao Hu, Li Chen, Penghao Wu, Hongyang Li, et al.
ECCV 2022
A spatial-temporal feature learning scheme towards a set of more representative features for perception, prediction and planning tasks simultaneously.

Delving into the Devils of Bird's-Eye-View Perception: A Review, Evaluation and Recipe

Hongyang Li, Chonghao Sima, Jifeng Dai, Wenhai Wang, Lewei Lu, et al.
arXiv 2022 [Setup the Table]
We review the most recent work on BEV perception and provide analysis of different solutions.

BEVFormer v2: Adapting Modern Image Backbones to Bird's-Eye-View Recognition via Perspective Supervision

Chenyu Yang, Xizhou Zhu, Hongyang Li, Jifeng Dai, et al.
CVPR 2023 Highlight
A novel bird's-eye-view (BEV) detector with perspective supervision, which converges faster and better suits modern image backbones.

Towards Capturing the Temporal Dynamics for Trajectory Prediction: a Coarse-to-Fine Approach

Xiaosong Jia, Li Chen, Penghao Wu, Jia Zeng, et al.
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.

HDGT: Heterogeneous driving graph transformer for multi-agent trajectory prediction via scene encoding

Xiaosong Jia, Penghao Wu, Li Chen, Hongyang Li, Yu Liu, Junchi Yan
arXiv 2022
HDGT formulates the driving scene as a heterogeneous graph with different types of nodes and edges.

Stare at What You See: Masked Image Modeling without Reconstruction

Hongwei Xue, Peng Gao, Hongyang Li, et al.
CVPR 2023
An efficient MIM paradigm MaskAlign and a Dynamic Alignment module to apply learnable alignment to tackle the problem of input inconsistency.