Rapid Development of Diverse Human Body Models for Crash Simulations through Mesh Morphing 2016-01-1491

Current finite element (FE) human body models (HBMs) generally only represent young and mid-size male occupants and do not account for body shape and composition variations among the population. Because it generally takes several years to build a whole-body HBM, a method to rapidly develop HBMs with a wide range of human attributes (size, age, obesity level, etc.) is critically needed. Therefore, the objective of this study was to evaluate the feasibility of using a mesh morphing method to rapidly generate skeleton and whole-body HBMs based on statistical geometry targets developed previously.

THUMS V4.01 mid-size male model jointly developed by Toyota Motor Corporation and Toyota Central R&D Labs was used in this study as the baseline HBM to be morphed. Radial basis function (RBF) was used to morph the baseline model into the target geometries. The statistical skeleton geometry models (ribcage, femur, pelvis and tibia) and body shape model used in this study were previously developed based on CT images from more than 300 subjects and whole body scans from more than 200 subjects in total. Using the age, sex, height, and body mass index (BMI), the statistical skeleton model predicted the bone shape and cortical thickness, while the body shape model predicted the body surface contour and locations of the joints and anatomical surface landmarks.

RBF mesh morphing was conducted at the skeleton level as well as the whole-body level across a wide range of body sizes (5^th to 95^th percentile height for male) and shapes (BMI 20 to 40). The morphing process was automated, so only 2-3 hours were needed to conduct a whole-body mesh morphing on a personal computer. It was found that the morphed skeleton models generally sustained similar mesh qualities as those in the baseline model. The morphed whole-body HBMs may sustain lower mesh qualities in the areas whose geometries are very different to the baseline model. However, these reduced mesh qualities can be improved by adding a smoothing function or better controlling the morphing landmarks during the mesh morphing process.

This study demonstrated that the skeleton and body shape geometries can be rapidly predicted to represent a wide range of human characteristics. The mesh morphing method can rapidly develop skeleton and whole-body HBMs with different body sizes and shapes with a mesh quality similar to the baseline model.