Empirical Inference Conference Paper 2012

A Kernel-based Approach to Direct Action Perception

The direct perception of actions allows a robot to predict the afforded actions of observed novel objects. In addition to learning which actions are afforded, the robot must also learn to adapt its actions according to the object being manipulated. In this paper, we present a non-parametric approach to representing the affordance-bearing subparts of objects. This representation forms the basis of a kernel function for computing the similarity between different subparts. Using this kernel function, the robot can learn the required mappings to perform direct action perception. The proposed approach was successfully implemented on a real robot, which could then quickly learn to generalize grasping and pouring actions to novel objects.

Author(s): Kroemer, O. and Ugur, E. and Oztop, E. and Peters, J.
Book Title: International Conference on Robotics and Automation (ICRA 2012)
Pages: 2605--2610
Year: 2012
Month: May
Day: 0
Publisher: IEEE
Bibtex Type: Conference Paper (inproceedings)
DOI: 10.1109/ICRA.2012.6224957
Event Name: IEEE International Conference on Robotics and Automation (ICRA 2012)
Event Place: St. Paul, MN, USA
State: Published
Digital: 0
Electronic Archiving: grant_archive
Links:

BibTex

@inproceedings{KroemerUOP2012,
  title = {A Kernel-based Approach to Direct Action Perception},
  booktitle = {International Conference on Robotics and Automation (ICRA 2012)},
  abstract = {The direct perception of actions allows a robot to predict the afforded actions of observed novel objects. In addition
  to learning which actions are afforded, the robot must also learn to adapt its actions according to the object being manipulated. In this paper, we present a non-parametric approach to representing the affordance-bearing subparts of objects. This representation forms the basis of a kernel function for computing the similarity between different subparts. Using this kernel function, the robot can learn the required mappings to perform direct action perception. The proposed approach was successfully implemented on a real robot, which could then quickly learn to generalize
  grasping and pouring actions to novel objects.},
  pages = {2605--2610},
  publisher = {IEEE},
  month = may,
  year = {2012},
  slug = {kroemeruop2012},
  author = {Kroemer, O. and Ugur, E. and Oztop, E. and Peters, J.},
  month_numeric = {5}
}