Dirk Kraft

About

Dirk Kraft is a robotics and cognitive systems researcher whose work sits at the intersection of robot perception, manipulation, and autonomous learning. He is best known for pioneering the concept of **Object-Action Complexes (OACs)**, a framework for grounding abstract sensory-motor processes in robotic systems, which has garnered over 150 citations and become an influential reference in developmental robotics. His research has made substantial contributions to how robots discover, represent, and interact with objects in their environment — particularly through autonomous exploration rather than explicit programming. Kraft's work on grasp affordance learning is especially impactful. His papers on learning grasp affordance densities and refining these models through experience — together accumulating over 120 citations — introduced probabilistic, nonparametric approaches to modeling how robots can identify stable grasping configurations. His broader research program explores how robots can bootstrap rich object and action knowledge from scratch, mirroring aspects of infant cognitive development. Beyond manipulation, Kraft has contributed to natural language integration in humanoid robots and fast-setup automated assembly processes, reflecting a versatile research portfolio. His body of work is essential reading for anyone studying robot learning, affordance theory, or cognitive developmental robotics.

Research Focus

Key Achievements

Top Papers

1. 1
   Object–Action Complexes: Grounded abstractions of sensory–motor processes

   153 citations · 2011
2. 2
   BIRTH OF THE OBJECT: DETECTION OF OBJECTNESS AND EXTRACTION OF OBJECT SHAPE THROUGH OBJECT–ACTION COMPLEXES

   81 citations · 2008
3. 3
   Learning Grasp Affordance Densities

   80 citations · 2011
4. 4
   Refining grasp affordance models by experience

   40 citations · 2010
5. 5
   Development of Object and Grasping Knowledge by Robot Exploration

   36 citations · 2010
6. 6
   Early Reactive Grasping with Second Order 3D Feature Relations

   33 citations · 2007
7. 7
   Structural Bootstrapping—A Novel, Generative Mechanism for Faster and More Efficient Acquisition of Action-Knowledge

   27 citations · 2015
8. 8
   Technologies for the Fast Set-Up of Automated Assembly Processes

   24 citations · 2014
9. 9
   Integrating multi-purpose natural language understanding, robot’s memory, and symbolic planning for task execution in humanoid robots

   22 citations · 2017
10. 10
    Learning Objects and Grasp Affordances through Autonomous Exploration

    22 citations · 2009

Key Collaborators

Contact & Links