Welfare-oriented Interactive Interface based on the Model of Knowledge, Emotion and Intention

近年来，信息基础设施正在渗透到每个家庭，有效地利用这些基础设施实现智能接口是我们所期望的。在本研究中，我们的目标是在一个智能界面中实现两个重要的功能。1.系统不使用键盘、鼠标等，通过肢体表情或语言理解用户意图的功能。系统侧了解用户情绪(愤怒、痛苦等)的功能到去年为止，我们的研究分为(1)和(2)两个项目，分别进行了研究。(1)从肢体表情或语言的角度对指令手势进行识别、获取和修改的研究，从用户的情境中提取有意义的手势。从简单的肢体表情和语言中获取和修改指令符号，以执行两个或两个以上的指令。(2)从肢体表情和语言中检测情感和意图，并研究引导和支持的方法。此外，我们还开展了信息系统的指导方法和福利制度的支持方法的研究。今年，我们整合了上述(1)和(2)，并建立了一个实验系统。该系统能够从用户的身体表情中识别出有意义的手势。该系统还能够识别用户S的情感和意图，并利用知识、情感和意图的工程模型使用户得到有效的支持。我们能够展示出实现比传统福利支持系统更积极的支持的可能性。

In recent years, an information infrastructure is permeating each home, and it is desirable that we realize the intelligent interface utilizing the infrastructure effectively. In this research, we aim to realize two important functions in an intelligent interface. 1. Function that the system understands user's intentions with body expression or language, without using a keyboard, a mouse, and so on. 2.Function that system side understands user's emotion (anger, suffering, etc.) and carries out the user support by using the model of knowledge, emotion and intention.Until last year, we divided our research into the following items (1) and (2), and researched them individually.(1)Research about recognition, acquisition and modification of command signs from body expression or language.We carried out research for extracting meaningful signs from user's situations. We also carried outresearch for acquisition and modification of command signs in order to execute two or more commands from simple body expression or language.(2)Research about detection of emotion and intention from body expression and language, and research about method of guidance and support.We advanced research of the engineering model of knowledge, emotion and intention, in order to detect user's emotion and intention. Furthermore, we carried out research of a guidance method on information systems and a support method on welfare systems.This year, we integrated the above-mentioned (1) and (2), and built an experimental system. The system was able to recognize the meaningful signs from user's body expression. The system was also able to recognize user s emotion and intention, and enabled the user to be given the effective support for him by using the engineering model of knowledge, emotion and intention. We were able to show the possibility of realizing more positive support than conventional welfare support system.