基于认知思维的产品意象造型智能设计

发布时间：2018-03-25 12:30

本文选题：产品设计　切入点：用户认知　出处：《兰州理工大学》2014年博士论文

【摘要】：产品造型形态是表达产品设计思想与实现产品功能的语言和媒介,能够传达精神、文化等层面的意义与象征性。它是产品自身功能、结构、材料以及工艺技术等客观因素与设计师和用户在审美、价值判断等主观因素相互作用的综合结果。基于此,本研究针对产品意象造型智能设计相关问题,依据用户与设计师的认知思维规律,寻求进一步提升产品意象造型智能设计系统性能的方法以及设计思维品质评价的方法。本研究的主要研究内容与获得的结论如下。 (1)提出应用数量化理论建立符合人的视知觉选择分辨性和简约调节性的产品意象造型设计要素辨识模型。应用该模型在感性工学系统的理论架构内对产品造型设计要素进行分解,以造型设计要素为自变量、用户情感需求为因变量,探讨用户的感性意象与产品造型设计要素的对应关系,并据此辨识产品造型设计要素针对各感性意象的个性要素和平台要素。建立计算机辅助产品意象造型设计要素分析系统,将该系统所分析的产品造型设计要素数据应用于实际项目,取得了较好的效果。 (2)依据人的视知觉整体组织性、恒常记忆性、选择分辨性和简约调节性构建神经网络,提出了改良型四层BP神经网络。基于改良型四层BP神经网络建立以产品形态要素为输入、用户情感意象需求为输出、用户情感需求调查结果为训练样本的产品意象智能评价系统。与基于常规三层BP神经网络的产品意象智能评价系统进行对比研究,结果表明基于改良型四层BP神经网络的产品意象智能评价系统具有较高的预测精度。 (3)通过分析设计师的设计思维流程及其设计进程、设计思维模式与设计策略所构成的关系单元,提出了基于广义交互式遗传算法的设计师设计思维模型。应用基于精英保留策略的非劣分类遗传算法建立人机交互式产品多意象造型智能设计系统(IIDSPMIF)模拟产品设计过程中的多意象目标优化。实验结果表明,IIDSPMIF系统总体上符合设计师的设计流程,能够模拟设计师的设计策略和设计思维模式。 (4)提出设计思维品质评价方法,该方法应用思维数学的相关理论将产品设计过程转换为设计师的思维网络,采用系统套、思维坐标系等理论量化思维环节,求解设计思维品质函数。应用该方法分析了设计师之间的思维品质差异以验证其可行性。最后,应用该方法对比分析了广义设计师C-IIDSPMIF与设计师C的思维品质,结果表明IIDSPMIF系统有效提升了广义设计师的思维独创度、统摄度、流畅度等方面的思维品质。本研究所取得的研究成果为指导企业开展新产品的意象造型设计进行了有益的探索与尝试。另外,设计思维品质评价方法的建立为研究设计认知思维提供了新的实践与理论基础。
[Abstract]:Product modeling form is the language and medium to express product design idea and realize product function. It can convey the meaning and symbolism of spirit and culture. It is the function and structure of the product itself. The synthetic results of the interaction between the objective factors such as materials and process technology and the subjective factors such as aesthetics and value judgment of designers and users. Based on this, this study aims at the problems related to the intelligent design of product image modeling. According to the cognitive thinking law of users and designers, this paper seeks to further improve the performance of product image modeling intelligent design system and the evaluation method of design thinking quality. The main contents and conclusions of this study are as follows. 1) applying the quantitative theory to establish the identification model of product image modeling elements which accords with human visual perception choice resolution and minimalism adjustment. The model is applied to product modeling in the theoretical framework of perceptual engineering system. Factoring the elements, This paper discusses the corresponding relationship between the perceptual image of the user and the design elements of the product modeling, taking the styling design element as the independent variable and the user's emotional demand as the dependent variable. According to the identification of product modeling design elements, aiming at the individual elements and platform elements of each perceptual image, a computer aided product image modeling design element analysis system is established. The data of product modeling design elements analyzed by this system are applied to practical projects, and good results are obtained. (2) according to the overall organization of human visual perception, constant memory, selective resolution and minimalist adjustment, a modified four-layer BP neural network is proposed, which is based on the improved four-layer BP neural network. The product image intelligent evaluation system based on user's emotional image demand is output, and the result of user's emotional demand investigation is the product image intelligent evaluation system based on training sample, which is compared with the product image intelligent evaluation system based on conventional three-layer BP neural network. The results show that the product image intelligent evaluation system based on improved four-layer BP neural network has high prediction accuracy. 3) by analyzing the design thought flow and the design process of the designer, the relationship unit between the design thinking mode and the design strategy, This paper presents a designer's design thinking model based on generalized interactive genetic algorithm (GIGA), and applies the non-inferior classification genetic algorithm based on elite reservation strategy to establish the simulation product setup of human-computer interactive multi-image modeling intelligent design system (IIDSPMIF). The experimental results show that the IIDSPMIF system generally accords with the designer's design process. Can simulate the designer's design strategy and design thinking mode. (4) the evaluation method of design thinking quality is put forward, in which the product design process is transformed into the designer's thinking network by using the relevant theory of thinking mathematics, and the system set, thinking coordinate system and other theoretical quantification thinking links are adopted. The difference of thinking quality among designers is analyzed by using this method to verify its feasibility. Finally, the thinking quality of generalized designer C-IIDSPMIF and designer C is compared with that of designer C. The results show that the IIDSPMIF system can effectively improve the thinking quality of the generalized designers in terms of originality, dominance, fluency and so on. The research results obtained in this study have made a beneficial exploration and attempt to guide enterprises to carry out image modeling design of new products. The establishment of design thinking quality evaluation method provides a new practical and theoretical basis for the study of design cognitive thinking.
【学位授予单位】：兰州理工大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TB472

【参考文献】