Papers
16
Total Citations
498
H-Index
9
About
Bowen Wen is a robotics researcher whose work sits at the intersection of computer vision, 6D pose estimation, and dexterous robot manipulation. His most influential contributions center on enabling robots to reliably perceive and interact with objects in complex, real-world settings. His landmark systems, SE(3)-TrackNet (123 citations) and BundleTrack (116 citations), tackled the notoriously difficult problem of tracking object poses across video sequences — the former by cleverly leveraging synthetic training data, and the latter by eliminating the need for pre-existing 3D CAD models, dramatically broadening applicability to novel objects. These advances have become foundational references in robotics perception research. Beyond perception, Wen has made notable contributions to contact-rich manipulation, including compliant finger gaiting for in-hand dexterity (72 citations), high-precision assembly using vision-driven compliance (59 citations), and force-guided sim-to-real transfer through the FORGE framework. His HANDAL dataset further supports the broader community by providing richly annotated, manipulation-focused object data including affordances and reconstructions. With over 470 cumulative citations spanning perception, planning, and manipulation, Wen has established himself as a versatile and impactful contributor to modern robot autonomy research.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10