基于眼动追踪的地铁客室设计意象评价  

Image evaluation of subway passenger room design based on eye tracking

作　　者：李艳 李欢 姜彦廷 宋武 李鹏飞 LI Yan;LI Huan;JIANG Yanting;SONG Wu;LI Pengfei(College of Mechatronic and Automation,Huaqiao University,Xiamen 361021;CRRC Zhuzhou Electric Locomotive Co.,Ltd.,Zhuzhou 412000)

机构地区：[1]华侨大学机电及自动化学院,福建厦门361021 [2]中车株洲电力机车有限公司,湖南株洲412000

出　　处：《机械设计》2023年第5期150-155,共6页Journal of Machine Design

基　　金：福建省自然科学基金(2019J01061);2020年度教育部人文社会科学研究项目(20YJA760067);中车株洲电力机车有限公司与华侨大学合作研究项目(2021-0015)。

摘　　要：以提升地铁乘客的乘车体验为目标,采用眼动追踪技术和BP神经网络算法构建了地铁客室设计评价模型,为产品改进设计提供目标和方向。以不同主题风格的地铁客室为素材,通过Tobii X3-120眼动仪采集30位乘客观察地铁客室不同兴趣区的眼动数据作为BP算法的输入,将用户对意象的评分作为输出,在眼动数据与意象评分间建立两者关联的数学模型。模型经测试后,得出地铁设计意象评价预测值与真实值的绝对平均误差为0.8,结果表明:该神经网络能较准确地预测乘客对地铁客室设计的意象评分,验证了模型的有效性。将眼动指标融入人工神经网络算法,相对意象评价中采用访谈及问卷的方法,更为全面与客观,为今后地铁客室设计意象的评价与设计提供参考。Aiming at improving subway passengers ride experiences,a design evaluation model of subway passenger room intention was established combined with eye tracking technology and BP neural network algorithm,providing objectives and directions for design optimization.Taking subway passenger rooms with different theme styles as experimental materials,Tobii X3-120 eye tracker was used to collect the eye movement data of 30 passengers observing different areas of interest in the subway passenger rooms.These eye movement data was used as the input of BP algorithm,and the use image score was the output.The mathematical correlation model between eye movement data and image score was established.The model was tested and the absolute average error between the predicted value and the true value of the subway design image evaluation was 0.8.The prediction chart showed that the neural network could accurately predict the image scores of passengers on the subway passenger room design,which verified the validity of the model.Eye movement index was integrated into artificial neural network algorithm.Compared with the interview and questionnaire methods used in the previous image evaluation,the evaluation method was more comprehensive and objective,which could provide references for the evaluation and design of subway passenger room design image in the future.

关 键 词：感性工学 设计意象评价 眼动追踪 地铁客室设计 人工神经网络 

分 类 号：TB472[一般工业技术—工业设计]