Guest Editorial: Affective Robotics

ADVANCES in generative AI and machine learning are heralding a new era in sensing, interpreting, simulating, as well as stimulating and sometimes mitigating human emotion. While affective computing has traditionally focused on these processes within digital or screen-based interfaces, a new frontier is emerging where these capabilities are embodied in physical agents. Building affective robots remains one of the most demanding goals at the intersection of affective computing and robotics. It requires a unified view of approaches inspired by neuroscience, psychology, cognitive science, developmental and behavioral science, artificial intelligence, and robotics. Although current state-of-the-art robots can recognize and mimic simple emotions, the field has yet to produce such a genuinely affective robot: an artificial agent whose affect is not only perceived externally, but also represented internally, regulated over time, and meaningfully coupled to perception, action, and social understanding. At the same time, engineering robot emotions is also a scientific probe: it can sharpen hypotheses about how affect emerges in humans and other animals, by forcing theories to become computational, embodied, and testable in the wild. This special issue of the IEEE Transactions on Affective Computing seeks to address this gap by offering a prestigious venue for the discussion of affective robotics.