Towards a model for automatic action recognition for social robot companions

Ingo Keller, Markus Schmuck, Katrin Solveig Lohan

Research output: Chapter in Book/Report/Conference proceedingConference contribution

34 Downloads (Pure)

Abstract

In this paper, we will explore human movementusing a tutoring spotter system which controls an iCub robot.We will present an evaluation based on the captured humanmovement from an experimental study, where our participantsdemonstrated a salt-shaker and a cup-stacking task to the iCubrobot. We will use a method of action recognition, which willhelp the robot to differentiate between these actions, as it willfocus the robot’s attention on the vital information presented bythe human. Our findings imply that the behaviour of the robotaffects our participants and it influences both, presentationtime as well as the ratio between action and sub-action duringthe task presentation. Furthermore, the stability of the actionrecognition system is influenced by this modification of thehuman presentation.
Original languageEnglish
Title of host publication2016 25th IEEE International Symposium on Robot and Human Interactive Communication (RO-MAN)
PublisherIEEE
Pages85-90
Number of pages6
ISBN (Electronic)9781509039296
DOIs
Publication statusPublished - 17 Nov 2016
Event25th IEEE International Symposium on Robot and Human Interactive Communication 2016 - New York, United States
Duration: 26 Aug 201631 Aug 2016

Publication series

NameIEEE International Symposium on Robot and Human Interactive Communication
PublisherIEEEE
ISSN (Electronic)1944-9437

Conference

Conference25th IEEE International Symposium on Robot and Human Interactive Communication 2016
Abbreviated titleRO-MAN 2016
Country/TerritoryUnited States
CityNew York
Period26/08/1631/08/16

Fingerprint

Dive into the research topics of 'Towards a model for automatic action recognition for social robot companions'. Together they form a unique fingerprint.

Cite this