Home /Research /Multi-Level Human Robot Interaction by Aligning Different Ontologies
SWARM

Multi-Level Human Robot Interaction by Aligning Different Ontologies

Shotaro Kobayashi, Takahira Yamaguchi

Year
2012
Citations
2

Abstract

This paper discusses how to align different kinds of ontologies: Japanese Wikipedia Ontology, Behavior Ontology and Robot Kinematics Ontology with built-in software, in order to design Multi-Level HRI (Human Robot Interaction). The multi-level interaction includes three-way interactions: knowledge-level interaction with dialog and fact type QA, behavior-level interaction between a user and a humanoid robot, and task-level interaction with multi robot coordination to do a task given from a user. Japanese Wikipedia Ontology works for language-level interaction, Behavior Ontology works for bridging the gaps between human behavior and robot one. Robot behavior goes well with Robot Kinematics. More complicated interaction can be done where three-way interactions invoke each other. Case studies show us that children in elementary school and elder persons enjoy the Multi-Level HRI.

Keywords

Human–computer interactionHuman–robot interactionRobotComputer scienceArtificial intelligence

Related papers

Browse all SWARM papers