Researchers at the University of Glasgow's School of Computing Science have studied how faces and voices affect the emotional appeal of social robots. Dr Shaun Macdonald and Josh Yip used augmented reality to overlay dozens of face and voice combinations on Qoobo, an existing social robot with a soft, cushion-like body and a swinging tail. Macdonald said the aesthetic design of zoomorphic robots has so far largely been educated guesswork.
The team built five prototype faces: animal-like, robot-like, emoji-style, anime-inspired, and one made only of eyes. They paired them with five vocalisations: cat noises, human-like vocal sounds, abstract electronic noises, music, and 'animalese', the nonsense speech from Nintendo's Animal Crossing. Twenty-four volunteers wearing AR headsets ranked combinations for how clearly they expressed emotion, how much empathy they evoked and how appropriate they felt. The work builds on Macdonald's AZRA software, which overlays AR sounds and animations on real robots.
Animal-like faces and voices that matched Qoobo's animal-like body were well liked, but anime-style faces ranked just as highly. Yip said anime faces are more expressive than animal faces could be without feeling uncanny. Faces consistently outperformed sounds at conveying emotion; human sounds were easier to read emotionally but liked less than animal sounds. Animalese ranked highly despite being unintelligible, with some users feeling it suggested greater intelligence.
Although participants said correctly reading the robot's emotions was very important, the data showed readability had little influence on the designs they preferred. Pet owners were more likely to treat animal-like prototypes as real pets and robot-like faces as machines. Macdonald said no single face or voice has universal appeal, but the combination strongly affects how well users feel robots communicate emotion.
The researchers say the findings could inform future zoomorphic robots and enable user-customisable designs, and could extend the lifespan of existing devices by letting owners overlay new faces, voices and functions via AR. The paper, 'Anime, Animal or Animal Crossing? Comparing Aesthetic Styles for Zoomorphic Robot Faces and Sounds using Augmented Reality', was presented at the ACM MobileHCI conference in Swansea, with funding from the EPSRC and the University of Glasgow's EPSRC Impact Acceleration Account.