Evaluation of parametric sedan wheel hub based on Kansei Engineering and regression analysis

Abstract

<p><strong>Purpose:</strong> Parametric design has become one of the most important means of sedan design. The purpose of this paper is to construct a linear regression model to explore the matching relationship between Kansei words and the morphological elements of sedan wheels and to incorporate perceptual factors into the rational thinking of parametric wheel hub design. <strong>Design/methodology/approach:</strong> To forecast the matching relationship between wheel shape design features and semantics, this research offers a multiple linear regression model. First, using sample similarity matrix data analysis and offline research, 20 typical samples of car wheels were collected. Third, the wheel shape design elements are obtained by applying the semantic difference (SD) method to four groups of users’ evaluation data on the perceived words of the car wheels. Firstly, the wheel shape is divided into six basic features and seventeen subdivided features using morphological analysis. The collected car wheel samples are then researched and coded. Ultimately, a multiple linear regression model is built utilizing shape coding matching to the styles to direct the wheel parametric design process, yielding sedan wheel design schemes with various semantics of style. <strong>Findings:</strong> The results show that the regression models can provide good prediction performance (R² values are greater than 0.7). This study shows that the use of multiple regression models can accurately and cost-effectively predict the wheel hub morphological elements that meet the user’s perceptual needs through morphological elements and Kansei words. <strong>Originality/value:</strong> This paper is the first to use a multiple regression model to predict a parametric wheel shape that fits the user’s sensibility, which helps the user to present and feel the design scheme in a digital environment. The different wheel hub schemes generated by parametric design can be virtually displayed in the metaverse, helping users and designers to carry out more convenient scheme selections.</p>