感性驱动下的破拆机器人造型进化设计方法
摘 要：为提升破拆机器人造型质量，辅助设计者创造符合市场感性期望的外观，探究了一种基于交互式遗传算法的产品意象造型进化设计方法。运用贝塞尔曲线公式对造型特征面、线和点参数化，将产品的几何、功能和工艺等约束要求映射至数学模型上，并对特征线进行设计属性语义描述与归纳。通过数量化Ⅰ类理论分别构建特征线和点坐标对感性意象的多元线性方程，以决定系数作为特征线的感性权重，将两者串联实现造型预测。最后，应用该模型作为适应度函数结合交互式监督评价构建进化设计系统，根据实验结果，证明了该方法的可行性与合理性。
关键词：感性工学；交互式遗传算法；破拆机器人；曲线描述；造型量化
中图分类号：TB472 文献标识码：A 文章编号：1672-7053(2023)11-0076-04
Abstract：In order to improve the quality of the design of the demolition robot and assist designers in creating an appearance that meets the emotional expectations of the market, an evolutionary design method for product image modeling based on interactive genetic algorithm was explored. Using the Bezier curve formula to parameterize the shape feature surfaces, lines, and points, mapping the geometric, functional, and technological constraints of the product to the mathematical model, and describing and inducing the design attribute semantics of the feature lines. By quantifying Class I theory, we construct multiple linear equations of feature lines and point coordinates for perceptual images, and use the coefficient of determination as the perceptual weight of the feature lines. We concatenate the two to achieve shape prediction. Finally, the model was applied as a fitness function combined with interactive supervised evaluation to construct an evolutionary design system. Based on experimental results, the feasibility and rationality of this method were demonstrated.
Key Words：Kansei Engineering; Interactive Genetic Algorithm; Demolition Robot; Curve Description; Styling Quantification
