汽车工程 ›› 2023, Vol. 45 ›› Issue (8): 1479-1488.doi: 10.19562/j.chinasae.qcgc.2023.08.018

所属专题: 智能网联汽车技术专题-感知&HMI&测评2023年

• • 上一篇    下一篇

基于BP神经网络优化遗传算法的智能座舱感性意象预测

陈国强1,申正义1(),孙利2,支梦帆2,李彤2   

  1. 1.燕山大学机械工程学院,秦皇岛 066004
    2.燕山大学艺术与设计学院,秦皇岛 066004
  • 收稿日期:2023-03-08 修回日期:2023-06-03 出版日期:2023-08-25 发布日期:2023-08-17
  • 通讯作者: 申正义 E-mail:szy1317@foxmail.com
  • 基金资助:
    国家社科基金艺术学项目(21BG125)

Intelligent Cockpit Perceptual Image Prediction Based on BP Neural Network Optimization Genetic Algorithm

Guoqiang Chen1,Zhengyi Shen1(),Li Sun2,Mengfan Zhi2,Tong Li2   

  1. 1.School of Mechanical Engineering,Yanshan University,Qinhuangdao 066004
    2.School of Art and Design,Yanshan University,Qinhuangdao 066004
  • Received:2023-03-08 Revised:2023-06-03 Online:2023-08-25 Published:2023-08-17
  • Contact: Zhengyi Shen E-mail:szy1317@foxmail.com

摘要:

为减少主观因素干扰,满足用户多样化感性需求,提出基于BP神经网络优化遗传算法的智能座舱感性意象设计方法。从用户角度出发,获取用户感性意象并划分强度,应用因子分析法降维得到目标意象,获取新能源汽车座舱样本,应用聚类分析法筛选得到优势样本,并结合形态分析法提取智能座舱中控造型特征因子。基于BP神经网络构建特定目标意象与造型特征因子的映射模型,得到两者函数关系后将其作为适应度函数展开遗传算法分析,优选特定意象下造型因子的最优组合,完成评估方法与优选方法的结合。根据优势因子组合进行设计实践以验证方法实用性,结果表明该方法能有效满足用户多维感性需求,为智能座舱造型设计多样化提供新的思路和参考。

关键词: 智能座舱造型优化, 感性意象预测, BP神经网络, 遗传算法

Abstract:

In order to reduce subjective interference and meet the diverse emotional needs of users, the design method of intelligent cockpit perceptual image prediction based on BP neural network optimization genetic algorithm is proposed. From the user's point of view, user emotional image is obtained and the intensity is divided. Factor analysis method is used to reduce dimension to obtain target images and cockpit samples of new energy vehicles. Cluster analysis method is applied to screen and obtain advantage samples and the modeling characteristic factors of intelligent cockpit central control are extracted by combining with morphological analysis method. Based on BP neural network, the mapping model of specific target image and modeling feature factors is constructed, and the functional relationship between the two is obtained, which is used as fitness function to carry out genetic algorithm analysis, optimize the optimal combination of modeling factors under the specific image, and complete the combination of evaluation method and optimization method. According to the combination of advantage factors, the design practice is carried out to verify the practicability of the method. The results show that the method can effectively meet the multidimensional emotional needs of users, and provide a new idea and reference for the diversification of intelligent cockpit modeling design.

Key words: intelligent cockpit modeling optimization, perceptual image prediction, BP neural network, genetic algorithm