欢迎访问《图学学报》 分享到:

图学学报

• 工业设计 • 上一篇    下一篇

基于BP 神经网络的壁挂式充电桩产品形态研究

  

  1. 燕山大学艺术与设计学院,河北 秦皇岛 066004
  • 出版日期:2017-12-30 发布日期:2018-01-11
  • 基金资助:
    国家自然科学基金项目(51675464)

Research on Form of Wall Set Charging Pile Base on BP Neural Network

  1. School of Industrial Design, Yanshan University, Qinhuangdao Hebei 066004, China
  • Online:2017-12-30 Published:2018-01-11

摘要: 通过对壁挂式充电桩感性意象与形态设计要素的分析,建立两者间对应的BP 神经
网络设计模型,从而得到各感性意象对应的产品形态。利用感性工学中的李克特量法对收集到的
壁挂式充电桩形态样本和感性词汇进行设计评价并得到感性评价结果,将形态设计要素作为BP
神经网络模型的输入层参数,感性评价结果数值作为输出层参数,利用Matlab 软件进行壁挂式
充电桩BP 神经网络模型的反复训练和测试,最终得到壁挂式充电桩形态设计要素BP 神经网络
模型。最后通过MSE 度量该BP 神经网络模型准确性,研究表明建立的壁挂式充电桩感性意象和
形态设计要素间的BP 神经网络模型具有可行性。

关键词: 感性工学, BP 神经网络, 壁挂式充电桩, 造型设计

Abstract: By analyzing the sentimental images and elements of form of wall set charging pile, a BP
neural network design model is established between them. Use the method of Likert scale to evaluate
the form samples and sentimental words of wall set charging pile, calculate the matrix of sentimental
evaluation. Then encode the elements of form and input them into the BP neural network as input
parameters, normalize the data of matrix of sentimental evaluation as the output parameters. Finally
train and test the BP neural network of wall set charging pile by Matlab and establish the mapping
relation between elements of form and sentimental images. The BP neural network model of wall set
charging pile was obtained. The accuracy of BP neural network model is measured by the function of
mean square error. The results show that the BP neural network model between the sentimental
images and the elements of form of the wall set charging pile is feasible.

Key words: kansei engineering, BP neural network, wall set charging pile, modality design elements