The best available anthropometry data for U.S. children were gathered by an UMTRI team during the 1970s. These data have limited utility for vehicle applications because the standard anthropometric postures are dissimilar in important ways from the postures children use in vehicles. We conducted a study of 68 children from 40 to 100 lb sitting in a variety of vehicle- and booster-seat conditions to quantify the children's postures. We gathered data on torso dimensions, particularly depths and breathds, to provide more relevant data for crash dummy design. In general, children's vehicle-seated postures are more slumped than the fully erect postures measured in the standard anthro surveys.
We have extensive experience using the data from the 1977 UMTRI child anthropometry study conducted for the Consumer Product Safety Commission. The data are available on the downloads page, but must be used with care because they are not representative of contemporary children. Most noteably, U.S. children today are on average heavier at any particular age or stature, and hence weight-related measures, such as circumferences, are generally larger than reported in the 1970s data.
We have conducted detailed investigations of child body shapes for application to crash dummy design. In a recent study, we used a handheld laser scanner to record the shapes of children's spines and shoulders in a wide range of postures. We have used these data to develop target specifications for the corresponding parts of crash dummies and computational models representing children from ages 6 to 12. The shoulder data were analyzed using statistical methods similar to those applied to modeling skeletal geometry and whole-body shape data.
The surface point-cloud data were cleaned using semi-automated procedures in Geomagic Studio software. The resulting meshes were decimated to about 10k polygons and imported to Mathematica. Custom software was used to establish homologous meshes with respect to the underlying skeletal structures. These meshes were analyzed using principal component analysis and related methods to create models to predict average surface shape as a function of overall body dimensions, such as stature, sitting height, and biacromial breadth, or simpler measures like age.
The resulting model has been used to develop improved computational models of child occupants and to provide guidance for the development of advanced crash dummies. The body shape modeling approach is now being applied to whole-body surface modeling.
For more on anthropometric research, see:
©2013 Matthew P. Reed and The University of Michigan