A parametric design framework for the mass customization of bicycle helmet

Abstract

Cluster analysis of 3D head shapes plays a crucial role in the mass customization design of products related to the head. Head shapes exhibit variations across different races, and designing helmets exclusively for Chinese individuals cannot solely rely on or reference foreign head models. Currently, research on cluster analysis of Chinese head shapes is limited, especially concerning shape variances. To address this, we developed an improved k-medoids algorithm and integrated Cluster Validity Index as an assessment metric. This enabled us to cluster 339 Chinese young males aged 18 to 30 into 7 groups based on their head shapes. By comparing our improved algorithm to the traditional k-medoids method, we affirmed its superiority in achieving higher sample participation rates and reducing inter-cluster sample disparities. To simplify the helmet design and editing process, and to improve the efficiency of mass customization, we have developed a parametric modeling program for bicycle helmets based on the head shape clustering results. Results from the Helmet Fit Index and stress simulation analysis demonstrate that our approach significantly enhances helmet fit and wearer comfort.

Keywords

• 3D head shape clustering
• K-medoids algorithm
• Bicycle helmet
• Mass customization design
• Parametric modeling