Optimization and Minimization of Boundary Mannequins

Boundary mannequin is an important concept in digital human modeling and simulation, yet complicated to deal with and utilize. In theory, the number of boundary mannequins could be as much as (n!)2ⁿ for a single gender, where n is the number of critical anthropometric dimensions. It has been recommended ^[1] to break a complicated task into smaller tasks to reduce the scale of problem, and limit n=2 whenever possible. Even then, the number of boundary mannequins is still high for simulations. In this paper, the authors intend to further simplify the issue. An Excel worksheet is created for the purpose. The input can be as few as two points. An ellipse representing the boundary is automatically generated through regression analysis, and the extremes on the major and minor axes of the ellipse are then obtained, and taken as the optimal boundary mannequins. By doing so, the number of boundary mannequins is reduced to no more than four, and in many cases just one, and the process is standardized. The result is compared to the commonly used percentile mannequin. It is demonstrated that the optimal boundary mannequin is more reliable and safe to use than the percentile mannequin, but a carefully chosen percentile mannequin could be used to substitute the optimal boundary mannequin in preliminary tests to save time and cost.