Mapping Facial Expression to Internal States Based on Intuitive Parenting

Abstract

Sympathy is a key issue in interaction and communication between robots and their users. In developmental psychology, intuitive parenting is considered the maternal scaffolding upon which children develop sympathy when caregivers mimic or exaggerate the child’s emotional facial expressions [1]. We model human intuitive parenting using a robot that associates a caregiver’s mimicked or exaggerated facial expressions with the robot’s internal state to learn a sympathetic response. The internal state space and facial expressions are defined using psychological studies and change dynamically in response to external stimuli. After learning, the robot responds to the caregiver’s internal state by observing human facial expressions. The robot then expresses its own internal state facially if synchronization evokes a response to the caregiver’s internal state.