Synthesizing android and cognitive brain-scientific understanding of embodiment and sociality

综合机器人和认知大脑——对具身性和社交性的科学理解

基本信息

批准号：
15017214
负责人：
KUNIYOSHI Yasuo
金额：
$ 47.74万
依托单位：
The University of Tokyo
依托单位国家：
日本
项目类别：
Grant-in-Aid for Scientific Research on Priority Areas
财政年份：
2003
资助国家：
日本
起止时间：
2003 至 2005
项目状态：
已结题

项目摘要

In order to construct a design methodology for intelligent information systems with high affinity to humans, it is crucial to understand the basic principles of human cognition and behavior, that are, embodiment and sociality.In this research, three methods are used: (1) behavior/learning experiments using humanoid robots: (2) cognitive neuroscience based on brain activity measurements: and (3) experiments on human-robot interaction using " androids", new type of robots with perfectly human-like appearance. Through these experiments, the principles of human cognition such as embodiment and sociality are revealed from the three view points: robot, human, and the observer. The following results are obtained:1. Principles of action generation/recognition based on embodiment: Using an adult-size humanoid robot, the dynamic " roll-and-rise" action was realized for the first time in the world. This revealed the effect of embodiment on autonomous behavior. The same principles are also shown t … More o be in effect for action recognition.2. Development of androids with perfectly human-like outlook and their application to interaction experiments: In order to investigate a fundamental factor of human social interaction, namely the effect of outlook, androids with perfectly human-like outlook have been developed and demonstrated to the public. Furthermore, methodologies for realizing natural bodily movements for interactive dialogues have been developed.3. Principles and models of interaction: Brain activity measurements, looking time measurements and gaze measurements have been carried out on human subjects in dialogue situations involving either of human, android, non-humanoid robot, or simple machine. The experiments revealed how social cognition and attitude of the subjects change and develop for different targets. Also, a learning computational model which integrates autonomous embodied behavior and basic sociality, that is, imitation capability, was constructed.The above achievements have received outstanding international evaluation for the international publications and conference presentations. Furthermore, parts of the achievements have been featured in TV reports and newspapers, widely provoking interests from the public. Less

为了构建与人类高度亲和力的智能信息系统的设计方法论，理解人类认知和行为的基本原理，即体现性和社会性至关重要。在本研究中，使用了三种方法：（1）使用类人机器人进行行为/学习实验：（2）基于大脑活动测量的认知神经科学：以及（3）使用“机器人”进行人机交互实验，“机器人”是一种具有完美人类外观的新型机器人。人类认知的原则，例如具体化从机器人、人类和观察者三个角度揭示了其社会性，得到了以下结果： 1．世界上首次实现了“上升”动作。这揭示了具体化对自主行为的影响。同样的原理也适用于动作识别。2.开发具有完全人类特征的机器人像 Outlook 及其在交互实验中的应用：按顺序为了研究人类社会互动的一个基本因素，即外观的影响，已经开发出具有完全类似人类外观的机器人并向公众展示。此外，还开发了实现交互式对话的自然身体运动的方法。3．和交互模型：在涉及人类、机器人、非人形机器人或简单机器的对话中，对人类受试者进行了大脑活动测量、注视时间测量和凝视测量。这些实验揭示了社会认知和态度如何。主题根据不同的目标而变化和发展。并构建了融合自主体现行为和基本社会性即模仿能力的学习计算模型。上述成果在国际出版物和会议报告中均获得了优秀的国际评价。此外，部分成果已在电视报道和报纸，广泛引起公众的兴趣较少。