Endowing a NAO Robot With Practical Social-Touch Perception

Area of Science:

• Robotics
• Human-Computer Interaction
• Materials Science

Background:

• Socially assistive robots lack sophisticated touch-perception capabilities, limiting natural human-robot interaction.
• Augmenting existing robots with external touch systems offers a practical solution for enhanced social touch.
• Current robotic systems require advancements in tactile sensing for richer social interactions.

Purpose of the Study:

• To develop and evaluate a low-cost, easy-to-build soft tactile-perception system for the NAO robot.
• To assess the system's ability to detect affective touch gestures and force intensities.
• To gather user feedback on the tactile experience and its impact on human-robot interaction.

Main Methods:

• Fabric-and-foam-based resistive sensors were integrated onto the NAO robot's arm.
• Fifteen adults performed five types of affective touch gestures at two force intensities on the sensors.
• A random forest algorithm was trained to classify touch gestures and force intensities from sensor data.

Main Results:

• The tactile-perception system achieved 74.1% accuracy in identifying combined touch gestures and force intensities.
• This accuracy is over eight times better than chance, demonstrating effective touch detection.
• Participants found the sensor-equipped arm pleasant to touch and reported increased liking of the robot post-interaction.

Conclusions:

• The developed tactile-perception system can effectively detect social touch communication cues.
• The system is adaptable to various robot body parts, offering versatile application.
• This technology provides valuable tools for Human-Robot Interaction (HRI) researchers to implement social touch.