Multimodal Explainability in Child-Robot Interaction

摘要

Explainability is essential in child-robot interaction, shaping how children build trust, understanding, and confidence in robots.While explainability has been well studied for adult users, there is much less research focused on children, who need age-appropriate approaches.This research looks at how robots can use verbal, visual, and behavioural cues to explain their actions in ways that children can understand.The study is carried out in three stages.The first stage explores how different ways of giving information before the interaction affects how children see the robot.The second stage focuses on finding good explanation strategies during the interaction by using behavioural data and machine learning to help the robot adapt and respond in real-time.The third stage builds an autonomous robot that can explain its actions based on what it learns from the child's behaviour.This work will contribute to developing explainable robot behaviours that align with how children understand technology and provide guidance for designing trustworthy robots across different age groups and settings.