Emotion attribution to a non-humanoid robot in different social situations

Area of Science:

• Robotics and Human-Computer Interaction
• Ethology and Animal Behavior
• Artificial Intelligence and Emotion Recognition

Background:

• Growing interest in companion robots necessitates socially acceptable human-robot interaction (HRI).
• Meaningful HRI requires robots to convey intentionality and emotions for increased believability.
• Human-animal interaction, particularly with dogs, offers a valuable biological model for designing social robots due to their adaptive social behaviors.

Purpose of the Study:

• To design emotionally expressive robot behaviors inspired by canine behavior.
• To test human ability to recognize robot-expressed emotions and rule transgressions in an inter-specific context.
• To evaluate the effectiveness of dog-inspired behaviors in facilitating human attribution of emotional states to robots.

Main Methods:

• Observational study utilizing two wizard-of-oz experiments.
• Experiment 1: Assessed human recognition of robot-expressed "happiness" and "fear" and subsequent interaction.
• Experiment 2: Examined if humans could detect robot rule transgression based on greeting behavior.

Main Results:

• Humans readily attributed emotions to the social robot and interacted accordingly in Experiment 1.
• Participants successfully detected robot rule transgressions based on greeting behavior in Experiment 2.
• Dog-inspired behaviors proved effective in enabling humans to attribute emotional states to a non-humanoid robot.

Conclusions:

• Canine behavior provides a viable and effective model for designing emotionally expressive social robots.
• Humans are capable of recognizing and responding to robot-expressed emotions and rule-based behaviors.
• This approach enhances the believability and meaningfulness of human-robot interaction.