Margin: A Set of Lightweight Robots for Intervening in Child Dark Phobia via Embodied Interaction in Darkness

Dark phobia is a pervasive issue in early childhood, affecting both child development and family life. While the HCI field has explored exposure therapies via technologies such as Virtual Reality, there remains a lack of lightweight in-situ interventions tailored for domestic dark environments. To address this, we present Margin, a set of lightweight robots informed by principles of Ambient Intelligence and Embodied Interaction to support therapeutic play in domestic dark environments. The system comprises a guide robot, Heechu, integrating a large language model (LLM) for adaptive dialogue, and four ambient intervention robots operating in a coordinated multi-robot architecture. Through multisensory storytelling and interaction, Margin mediates children's engagement with dark spaces by reframing fear-related stimuli as approachable elements. This interaction paradigm enables children to actively participate in shaping their surroundings, supporting cognitive reframing of fear within a structured interactive context.