An Evolution Method of Driving Seat Comfort Based on Support Vector Regression

Sun Shouqian,Wu Qun,Wu Jianfeng,Shi Yuanwu,Yu Jianghong
DOI: https://doi.org/10.3321/j.issn:1004-132X.2008.11.015
2008-01-01
Abstract:An evaluation method based on support vector regression(SVR) was put forward for the purpose of predicting subjective perceptions of automobile seat comfort. The inputs included fourteen seat interface pressure measures, three anthropometric. The output was an overall comfort index derived from occupant responses to a survey. The SVR developed and validated using data collected from 12 subjects,and the subjects evaluated five different driving seats. The prediction results reach 0.0018 in mean square error and 0.869 in linear correlation coefficient; it gives better results than that with artificial neural network. The resulting knowledge should reduce the cost and time associated with the current automobile seat comfort development process.
What problem does this paper attempt to address?