Papers
9
Total Citations
489
H-Index
9
About
Dov Katz is a pioneering roboticist whose research lies at the intersection of interactive perception, autonomous manipulation, and learning in unstructured environments. His core contribution is the concept of "interactive perception"—the idea that robots should deliberately interact with their surroundings to gather sensory information, rather than passively observing. This paradigm shift enables robots to robustly segment, track, and model unknown objects, including articulated ones like doors and drawers, even when they are cluttered in piles. His most influential work, "Manipulating articulated objects with interactive perception" (139 citations), established a foundational approach for coupling manipulation and perception. Katz further advanced the field by developing methods for learning object affordances and grounded relational representations, allowing robots to autonomously acquire manipulation expertise from real-world interactions. His work on clearing piles of unknown objects and extracting kinematic models has been widely cited (over 370 total citations), demonstrating significant impact on autonomous robotics. Katz’s research directly addresses the fundamental challenges of object segmentation, action selection, and motion generation, making him a key figure in enabling robots to operate effectively in the messy, unpredictable environments of the real world.
Research Focus
Key Achievements
Top Papers
- 1Manipulating articulated objects with interactive perception139 citations · 2008
- 2
- 3
- 4
- 5Clearing a pile of unknown objects using interactive perception37 citations · 2013
- 6
- 7
- 8Extracting Planar Kinematic Models Using Interactive Perception17 citations · 2008
- 9Clearing a Pile of Unknown Objects using Interactive Perception10 citations · 2012