Natural human-robot interaction based on modeling behavior and information exchange

我们进行了研究，以实现与人自然互动的机器人。我们开发了一个机器人行为的计算模型，利用机器人的身体位置和方向来产生“在一起”的感觉，该模型通过检测伴侣的注意力转移并适当调整其身体位置和方向来再现人们在共同注意情境下的近距离行为。此外，我们调查了机器人的语言速度的影响，我们发现，在机器人和参与者边走边说话的情况下，参与者更喜欢较慢的语言速度。最后，在一项分析用户与机器人互动的研究中，我们发现他们的社交行为通常与他们对其他人的行为相似，尽管他们将较少的道德责任归咎于机器人。这表明人们将机器人视为一种超越人类和无生命物体的新类别。

We conducted studies toward realizing a robot that naturally interact with people. We developed a computational model of robot behavior that leads to feelings of "being together" using the robot's body position and orientation, in which people's proxemic behavior in joint attention situations is reproduced by detecting a partner's attention shift and appropriately adjust its body position and orientation in establishing joint attention with the partner. Further, we investigated the influence of a robot's speech rate, in which we found that participants prefer slower speech rate in a situation where the robot and participants talk while walking. Finally, in a study for analyzing users interaction with a robot, we found that they often behave in a social way similar to what they do to other people, though they attributed less moral responsibility to the robot. It indicates that people perceive a robot as a new category beyond neither humans nor inanimate object.