Signal Processing & Artificial-intelligence (SPA) Laboratory
Research
1. AI for intelligent vehicles and robotics
Multi-modal 3D perception/Radar perception
Perception resilient to sensor failures
3D occupancy prediction/Online map construction
Motion prediction and planning
Localization using BEV perception
E2E autonomous driving
Human action detection
2. Machine learning for generalization, robustness, and data efficiency
Multi-modal LLM-driven autonomous driving
Open-vocabulary semantic segmentation
Open-vocabulary 3D perception
Active learning
Long-tailed problem
Domain adaptation /generalization
Camera-Lidar sensor fusion 3D-CVF (trained not to detect truck or van) (ECCV 2020), code available: https://github.com/rasd3/3D-CVF
KITTI 3D object detection leaderboard (ranked 4th as of March 2020)
Joint 3D object detection and tracking (AAAI 2022)
Vedeo object detection (ICRA 2021)
3D object detection and tracking (trained not to detect truck or van)
LaPred Lane-aware trajectory prediction (CVPR 2021)
Vehicle trajectory prediction (IV2008)
Action detection on AVA dataset
Multimodal Human Action Recognition
