Papers
17
Total Citations
333
H-Index
11
About
Nicolas Pugeault is a researcher whose work spans computer vision, cognitive robotics, and machine learning, with particular expertise in how robots can autonomously learn to perceive and interact with their environments. His most influential contribution, "Birth of the Object" (2008, 81 citations), introduced a pioneering framework in which robots discover and segment objects through active exploration, linking visual and haptic perception within object-action complexes — a landmark step toward truly autonomous robotic cognition. Building on this, his work on grasping affordances and object representations demonstrated how robots could bootstrap rich world knowledge through self-directed exploration, without relying on pre-programmed knowledge. Pugeault also made significant contributions to visual representation, developing "visual primitives" — compact, semantically enriched local descriptors that improve robotic scene understanding while reducing computational overhead. His later research extended into contextual learning on humanoid robots using probabilistic models such as Latent Dirichlet Allocation, and semantic localization through deep learning, as seen in the SeDAR floorplan-reading system. Across his career, Pugeault has consistently advanced the frontier of intelligent robotic perception, producing work that bridges low-level visual processing with high-level cognitive understanding, making his research essential reading for those interested in autonomous robotics and embodied artificial intelligence.
Research Focus
Key Achievements
Top Papers
- 1
- 2Development of Object and Grasping Knowledge by Robot Exploration36 citations · 2010
- 3
- 4Early Reactive Grasping with Second Order 3D Feature Relations33 citations · 2007
- 5Learning Objects and Grasp Affordances through Autonomous Exploration22 citations · 2009
- 6
- 7
- 8
- 9
- 10