Construction of a Kansei Generative System for Product Form Design

[[abstract]]產品外形設計潛藏高度的不確定性與複雜性，而且相當不容易將之符碼參數化、公式化或法則化。在產品研發的前期階段中，現有的電腦輔助設計方法與系統並無法有效支援產品概念外形設計之相關活動。因此，本研究嘗試整合模糊理論、外形摻和法與遺傳基因演算法，利用產品意象知識庫，建立產品外形與感知意象評價之關連性，進以有效輔助產品外形衍生設計活動。 本研究為兩年期的目前僅針對單一組成元件與特徵之產品，進行外形衍生探討，後續的研究重點將針對多重特徵外形之產品，進行完整的自動化外形衍生系統建構。[[abstract]]The product form design activities involve a high degree of uncertainty and complexity and are therefore not easily formulated, coded and regularized. Consequently, very few of the computer-aided design approaches presented in the literature can support the conceptual form design tasks typically performed at the preliminary stages of a product’s development cycle. To enable designers to perform their design activities more objectively and efficiently, this paper combines the principles of fuzzy set theory, the shape-blending method and genetic algorithms to generate a knowledge-based approach for product form design based upon a database describing the relationships between different product forms and their corresponding perceptual image evaluations. The achievement of this two-year project currently conducts form generation for products with only one component of one feature. Subsequently, the research will aim at multi-featured product form for constructing the integrated system of automatic form generation.[[note]]NSC98-2221-E327-019-MY