Towards shared attention through geometric reasoning for Human Robot Interaction

摘要

Human robot interaction brings new challenges to the geometric reasoning and space sharing. The robot should not only reason on its own capacities but also consider the actual situation by looking from human's eyes, thus ¿putting itself to human's perspective¿. In humans, the ¿visual perspective taking¿ ability begins to appear by 24 months of age and is used to determine if another person can see an object or not. In this paper, we present a geometric reasoning mechanism that employs psychological concepts of ¿perspective taking¿ and ¿mental rotation¿ in order to reason what the human sees, what the robot sees and where the robot should focus to share human's attention. This geometric reasoning mechanism is demonstrated with HRP-2 humanoid robot in a human-robot face-to-face interaction context.