RESEARCH / 01
3D Scene Understanding
Perception, Mapping & Localization
We use cameras and LiDAR to understand 3D driving environments, studying object perception, mapping and localization, and automated data annotation.
- Camera & LiDAR-based Perception
- Automated Data Annotation
- SLAM-based Mapping & Localization
RESEARCH / 02
World Modeling
Scene Representation & Understanding
We represent the geometry, semantics, and temporal changes of driving environments observed by cameras and LiDAR in a unified world model. By combining 3D/4D scene representations with language-grounded understanding, we study environment modeling and simulation for autonomous driving.
- 3D/4D Scene Representation
- Spatiotemporal Environment Modeling
- Language-Grounded Scene Understanding
RESEARCH / 03
End-to-End Autonomous Driving
Policy Learning & Simulation
We study autonomous driving through end-to-end policy learning and simulation-based training.
- End-to-End Policy Learning
- Simulation-based Training