Collecting multi-modal data of human-robot interaction

Abstract

This paper presents a method and a tool to collect data of initiation stages in conversational interaction in order to better understand how humans talk to machines. This is done by means of a platform-robot which both facilitates conversations between a robot and human interlocutors and records the interaction using multi-modal technology. The system is provided with a standard face detection algorithm and an audio system and allows us to focus on the components of social interaction that make an interactive dialogue feel more natural.