Unsupervised learning of action primitives

Sanmohan, Volker Krüger, Danica Kragic

Research output: Chapter in Book/Report/Conference proceedingPaper in conference proceedingpeer-review

Abstract

Action representation is a key issue in imitation learning for humanoids. With the recent finding of mirror neurons there has been a growing interest in expressing actions as a combination meaningful subparts called primitives. Primitives could be thought of as an alphabet for the human actions. In this paper we observe that human actions and objects can be seen as being intertwined: we can interpret actions from the way the body parts are moving, but as well from how their effect on the involved object. While human movements can look vastly different even under minor changes in location, orientation and scale, the use of the object can provide a strong invariant for the detection of motion primitives. In this paper we propose an unsupervised learning approach for action primitives that makes use of the human movements as well as the object state changes. We group actions according to the changes they make to the object state space. Movements that produce the same state change in the object state space are classified to be instances of the same action primitive. This allows us to define action primitives as sets of movements where the movements of each primitive are connected through the object state change they induce.

Original languageEnglish
Title of host publication2010 10th IEEE-RAS International Conference on Humanoid Robots, Humanoids 2010
PublisherIEEE - Institute of Electrical and Electronics Engineers Inc.
Pages554-559
Number of pages6
ISBN (Electronic)9781424486908
ISBN (Print)9781424486885
DOIs
Publication statusPublished - 2010 Dec 1
Externally publishedYes
Event2010 10th IEEE-RAS International Conference on Humanoid Robots, Humanoids 2010 - Nashville, TN, United States
Duration: 2010 Dec 62010 Dec 8

Conference

Conference2010 10th IEEE-RAS International Conference on Humanoid Robots, Humanoids 2010
Country/TerritoryUnited States
CityNashville, TN
Period2010/12/062010/12/08

Subject classification (UKÄ)

  • Robotics

Fingerprint

Dive into the research topics of 'Unsupervised learning of action primitives'. Together they form a unique fingerprint.

Cite this