Papers
8
Total Citations
87
H-Index
5
About
Fuxin Li is a leading researcher in robotics and perception, whose work focuses on enabling robots to interact intelligently with unstructured and dynamic environments. His key research areas include robotic manipulation, grasp planning, and robust visual perception, with a particular emphasis on learning from physical interaction and 3D data. Li’s most impactful contribution is his pioneering work on learning contact locations for pushing and orienting unknown objects, which has garnered 45 citations. In this foundational study, he demonstrated how a robot can autonomously learn to predict effective push locations by experimenting with objects of varying shapes, using both local and global geometric features. This work laid the groundwork for subsequent advances in stable pushing and manipulation. Li has also made notable contributions to 3D data association with his GASP (Geometric Association with Surface Patches) framework, which robustly matches surface patches across views using only range data. More recently, his research has expanded into real-time generative grasping with spatio-temporal sparse convolution and video-based reasoning for occluded objects, addressing critical challenges in mobile and underwater manipulation. His 2024 work on point cloud models improving visual robustness in robotic learners highlights his ongoing commitment to bridging the gap between simulation and real-world deployment.
Research Focus
Key Achievements
Top Papers
- 1Learning contact locations for pushing and orienting unknown objects45 citations · 2013
- 2Learning stable pushing locations15 citations · 2013
- 3GASP: Geometric Association with Surface Patches11 citations · 2014
- 4Point Cloud Models Improve Visual Robustness in Robotic Learners6 citations · 2024
- 5Real-Time Generative Grasping with Spatio-temporal Sparse Convolution5 citations · 2023
- 6
- 7GASP : Geometric Association with Surface Patches2 citations · 2014
- 8