Autonomous Learning of Tool Affordances by a Robot

Abstract

This paper introduces a novel approach to representing and learning tool affordances by a robot. The tool representation described here uses a behavior-based approach to ground the tool affordances in the behavioral repertoire of the robot. The representation is learned during a behavioral babbling stage in which the robot randomly chooses different exploratory behaviors, applies them to the tool, and observes their effects on environmental objects. The paper shows how the autonomously learned affordance representation can be used to solve tool-using tasks by dynamically sequencing the exploratory behaviors based on their expected outcomes. The quality of the learned representation was tested on extension-of-reach tool-using tasks.

Download

Stoytchev, A., "Autonomous Learning of Tool Affordances by a Robot", In Proceedings of the Twentieth National Conference on Artificial Intelligence (AAAI), Pittsburgh, Pennsylvania, July 9-13, 2005..

A longer version of this paper can be downloaded here.

BibTeX Entry

@InProceedings{Stoytchev2005,
  author =     {Alexander Stoytchev},
  title =      {Autonomous Learning of Tool Affordances by a Robot},
  booktitle =  {Proceedings of the Twentieth National Conference on Artificial Intelligence (AAAI)},
  year =       2005,
  pages =      "??",
  address =    {Pittsburgh, Pennsylvania},
  month =      {July 9-13},
}

MPEG Movies from the Robot Experiments

Extension of Reach

The movies show testing trials for an extension of reach task with different tools.

L-Stick (1)
T-Stick (2)
T-Hook (3)

"Broken" Tool

Demonstration of autonomous adaptation under uncertainty. Learning is performed with a T-Hook. The movies show subsequent testing trials performed with an L-Hook. The robot "believes" it is still using the orignial T-Hook tool.