Research And Application Of Ontology-based Affective Design

In the current product design process, the consumers’ affective responses are used toestablish effective commucation bridge between designers and consumers, and affectivedesign has become more and more helpful for developing successful products. However, thereare still many problems that impede its intelligent support for product development. Kanseidata as an important part of affective design are lack of extensible structure. It is relativelyindependent in different design fields and is difficult for effective information sharing andreusing. Meanwhile, design concepts are defined according to the composition characteristicsof design objects, which cause great obstacles for knowledge processing. In addition, mostexpert system can only provide pre-existing samples in the database which reduces theeffectiveness of intelligent design support.In order to address these problems, we propose an ontology-based affective designsystem framework. This framework can make structured modeling for Kansei data, ensure theconsistency of these Kansei data in diverse design fields, improve the effectiveness ofinformation sharing and reusing, and then realize the automatic combination of designcandidates. We arrive at a new way of thinking in affective design by the formal descriptionof design concepts in ontology, and build the design knowledge base and combinationmechanism for design system.The main research works are shown as follows:(1) A framework for the affective design system is proposed. This framework covers thewhole process of design system establishment, including design concept building phase,ontology reasoning phase and design interaction phase. It can provide effective support foraffective design by extensible design knowledge base.(2) Design Object oriented ontology model is presented, and Kansei data are described into layered semantic concepts. Based on the general ontology model, we build the classmodel, property model and a detailed description of property constraints for card design andweb design. In addition, a theoretical method of updating and resuing domain ontology isdescribed for interdisciplinary data.(3) Design feature selection-combination mechanism for affective design is proposed.Emotional style is reasoned through different semantic layer and some seletion-combiantionmethods are applied to realize the automatic combination, including main element prioritymethod and top-to-down priority method. What’s more, the evaluation method for the designsystem is given.At last, a prototype system is developed for practical design field. Here, we use carddesign as an example. A Kansei evaluation expertiment is executed and its evaluation data areanalyzed by factor analysis and quantification theory type I method to achieve its Kansei database. Domain ontology is built based on hierarchical structure of general ontology modelwhich is used as a database in design system. In our affective design system, users can inputtheir affective demands and design contents, and then the system will provide an appropriatedesign scheme. Users can also do some personalized adjustment on the design scheme.Moreover, some experiments are conducted for the design schemes to verify the effectivenessof our affective design system.