Autonomous Achievement of Concept Structure for Event Recognition by a Robot

本研究的目标是实现对机器人上包含因果关系的动态事件的灵活解释能力和简报能力。本研究的一个具体目的是通过机器人自主地获取和结构化概念结构，我们提出了认知本体论的概念结构来解释机器人观察到的真实世界的事件。在本研究项目中，我们的研究目标分为以下四个方面：1)实现了快速识别颜色、几何图案、结构和运动轨迹的功能；2)实现了基于对象交互预测的目标关注对象的选择和更新技术；3)在实时观察的同时自动提取和获取概念结构的功能；4)基于机器人与被观察对象或事件之间的交互的认知本体论的主动确认功能。

The target of the present study is to achieve a flexible interpretation ability and the briefing ability to the dynamic event which is including the causality on the robot. A concrete purpose of this study is to acquire and to structurize the conceptual structure independently by the robot.We have proposed the conceptual structure "Cognitive Ontology" to interpret the event of the real world that the robot observed. In this research project, we divide our research target into the following four research topics:1) Realization of a function to execute recognition of color, geometry pattern, structure, and movement orbit in rapid way.2) Achievement of technology that robot can selects and renews a target object of attention for composes causal relation based on forecast of objects’ interaction.3) The function automatically to extract and to acquire conceptual structure while in real time observing it.4) Active acknowledgement function of Cognitive Ontology based on interaction between the robot and an observed objects or events.