Social Robotics - First Time Meeting With A Robot Using Backchannel Head Nodding
Author
Jensen, Daniel Grønkjær
Term
4. term
Education
Publication year
2014
Abstract
This thesis investigates human-robot interaction in first-time meetings where a robot listens and responds only with non-verbal backchannel head nods. It asks whether culturally appropriate feedback can help sustain face-to-face conversation and shape the speaker’s perception of the robot, and whether the robot’s embodiment influences these outcomes. A replication of an earlier study was conducted with Danish participants, complemented by pilot and main studies evaluating head-nod behaviors across different agent embodiments. The results contradict the original study, indicate that embodiment affects how the agent is perceived, and show that a humanoid robot using backchannel head nods is judged as more intelligent. The thesis discusses implications for designing autonomous listener systems that adapt to cultural norms.
Dette speciale undersøger menneske-robot interaktion i førstegangs-møder, hvor en robot lytter og kun reagerer med nonverbal backchannel-hovednik. Arbejdet spørger, om kulturelt passende feedback kan hjælpe med at fastholde en ansigt-til-ansigt samtale og påvirke talerens opfattelse af robotten, og om robotens kropslige/visuelle udformning (embodiment) har betydning for disse udfald. En replikation af et tidligere studie blev gennemført med danske deltagere og suppleret af pilot- og hovedstudier, der evaluerede hovednik på tværs af forskellige agent-udformninger. Resultaterne modsiger det oprindelige studie, indikerer at udformningen påvirker, hvordan agenten opfattes, og viser, at en humanoid robot med backchannel-hovednik vurderes som mere intelligent. Specialet drøfter implikationer for design af autonome lytter-systemer, der tilpasser sig kulturelle normer.
[This apstract has been generated with the help of AI directly from the project full text]