一种基于体压分布和PCA⁃BP神经网络模型的办公椅舒适度测定方法  

A method for measuring the comfort of office chair based on body pressure distribution and PCA-BP neural network model

作　　者：晁垚 金倩如 申黎明 谈立山 汪洋 CHAO Yao;JIN Qianru;SHEN Liming;TAN Lishan;WANG Yang(College of Furnishings and Industrial Design,Nanjing Forestry University,Nanjing210037,China;Anji Inspection&Testing Center,Huzhou 313300,Zhejiang,China)

机构地区：[1]南京林业大学家居与工业设计学院,南京210037 [2]安吉县质量技术监督检测中心,浙江湖州313300

出　　处：《林业工程学报》2021年第5期183-190,共8页Journal of Forestry Engineering

基　　金：浙江省质监系统质量技术基础建设项目(20180121)。

摘　　要：为了探索一种不依赖主观评价的办公椅使用舒适度测定方法,并建立一种可以直接用于产品评价以及设计研发辅助的办公椅使用舒适度评定模型,根据办公椅表面体压分布指标进行系统聚类,将26张实验椅聚类为5个类簇,综合分析各类簇办公椅表面体压分布指标的均值,将办公椅定义为5个舒适度等级。根据主成分分析法建立PCA-BP神经网络预测模型,使用表面体压分布指标预测办公椅舒适等级。结果表明:PCA-BP神经网络对各类簇办公椅平均预测误差为4.020%,对用于检测的3张办公椅预测结果符合办公椅表面体压分布特点,表明了该模型通过体压分布指标评定舒适度的可行性。通过系统聚类定义各类簇办公椅舒适度等级以及通过PCA-BP神经网络预测办公椅舒适度的方法是可行的,测定结果是可信的;该测定方法更加客观,能够避免主观评价随意性的缺陷;该模型使用时只需要测量办公椅表面体压分布指标值,可方便、高效应用在产品检测、评价以及新产品研发测试阶段,提高评价和设计效率。At present,the comfort evaluation of office chairs mainly relies on the method of combining objective measurement and subjective evaluation.In addition,the teacher signal of predicting comfort through neural network also relies on subjective evaluation.However,the subjective evaluation has the disadvantages of arbitrariness and inaccuracy,so the neural network that uses subjective evaluation value as teacher signal still has the previously mentioned disadvantages.Therefore,the significance of this study is to avoid the errors caused by the arbitrariness of subjective evaluation.In order to develop an office chair comfort evaluation method which does not depend on subjective evaluation and to establish a rapid comfort evaluation model which can be directly used in product evaluation and design,in this study,according to the pressure distribution index of the office chair surface,a systematic clustering was carried out,26 experimental chairs were divided into five categories,and office chairs were defined as five comfort levels by comprehensive analysis of the average of pressure distribution index of various office chairs.The principal component analysis(PCA)-BP neural network prediction model was established.The results indicate that the average prediction error of the PCA-BP neural network for all kinds of office chairs is 4.020%.The prediction results of the three office chairs can be used for testing accord with the characteristics of the pressure distribution on the seat surface,indicating the feasibility of the model to measure the comfort level through objective indicators.It is feasible to define the comfort level of various office chairs by system clustering and predict the comfort level of office chairs by the PCA-BP neural network,and the measured results are credible.The method is more objective and can avoid the defect of arbitrary subjective evaluation.This model can be conveniently and efficiently applied in product detection,evaluation,and product development,which only needs interface pressure,sav

关 键 词：办公椅 舒适度测定 体压分布 PCA-BP神经网络 聚类分析 

分 类 号：TB18[一般工业技术]