Incremental learning of full body motion primitives and their sequencing through human motion observation

Dana Kulić, Christian Ott, Dongheui Lee, Junichi Ishikawa, Yoshihiko Nakamura

Research output: Contribution to journalArticleResearchpeer-review

123 Citations (Scopus)

Abstract

In this paper we describe an approach for on-line, incremental learning of full body motion primitives from observation of human motion. The continuous observation sequence is first partitioned into motion segments, using stochastic segmentation. Next, motion segments are incrementally clustered and organized into a hierarchical tree structure representing the known motion primitives. Motion primitives are encoded using hidden Markov models, so that the same model can be used for both motion recognition and motion generation. At the same time, the temporal relationship between motion primitives is learned via the construction of a motion primitive graph. The motion primitive graph can then be used to construct motions, consisting of sequences of motion primitives. The approach is implemented and tested during on-line observation and on the IRT humanoid robot.

Original languageEnglish
Pages (from-to)330-345
Number of pages16
JournalInternational Journal of Robotics Research
Volume31
Issue number3
DOIs
Publication statusPublished - 1 Mar 2012
Externally publishedYes

Keywords

  • humanoid robots
  • learning by demonstration
  • motion primitive learning
  • stochastic models

Cite this