Human Body Models Customization by Advanced Mesh Morphing: Parametric THUMS

This work explores the feasibility of parametrizing the Total Human Model for Safety (THUMS) based on statistical anthropometric percentile using mesh morphing of THUMS AM50 driven by radial basis functions (RBF). The study first establishes the implementation criteria for mesh morphing, recognizing the modulation of shape differences between THUMS AM50 and AM95 as indicative of anthropometric variability between the generic percentile and the 50th percentile. The RBF problem is defined by selecting source points on well-distributed edges of the FE HBM and calculating displacement fields assigned to them. The mesh morphing process from the 50th to the generic percentile is fully automated and takes approximately 10 s. Shifting focus from modeling to verification, the study evaluates the effectiveness of the adopted mesh morphing strategy by comparing the shape of THUMS AM95 with that of its parametric counterpart AM50m95 (to indicate the 50th percentile transformed into the 95th) in a frontal sled test. Geometric verification confirmed close correspondence between THUMS AM95 and AM50m95, with a global average deviation of just 3.6 mm. Kinematic verification comparing the average displacements of specific landmark points is good at some locations but some detected deviations require further investigation. The parametric model is then used to span in the range 35–95 comparing AM50m35, AM50, AM50m75, and AM50m95 showing how percentile variation turns into an almost linear change of average displacement registered during the sled test.