Study on Multimodal System for Kansei Design Support

随着个人电脑和工作站的广泛使用，绘图软件支持的设计工作越来越受欢迎。但是，该软件并不辅助创作工作，只是为用户提供方便的绘图工具。我们的最终目标是实现艺术图案绘图的支持系统。我们分析了创造性工作的过程，并设计了支持和扩展用户创造性活动的用户界面。本文提出了感性设计支持系统的推理方法。首先，我们总结了推理方法不可缺少的特殊特征。其次，我们提出了一种新的推理方法，即基于模糊集的多阶段关联推理。第三，我们分析了所提出的推理机制的特点。最后通过实际设计过程验证了推理方法的有效性。

As the personal computers and the workstations are widely used, the design-work supported by the drawing software is getting popular. However, the software does not assist the creative work, but provide users with the convenient tool only for drawing. Our final goal is to realize the supporting systems for the artistic pattern drawings. We have analyzed the process of the creative work and designed the user interface that supports and expands the user's creative activities. In this paper, we propose the inference method for our kansei design support system. First we summarize the special features that are indispensable for the inference method. Second we propose the new inference method, called the Fuzzy Set-based Multistage Associative Inference, to cope with the requirements. Thirdly, we analyze the characteristics of the proposed inference mechanism. Finally we demonstrate the effectiveness of the inference method through the actual design processes.