A modern take on the theoretical modelling of inertial properties of a human body for biomechanical simulations

Abstract

The accurate estimation of inertial properties of individual body segments, including mass and moment of inertia, for biomechanical modelling and simulation is important in predicting realistic movement patterns. These properties must be subject-specific, as body shape and weight distribution varies significantly between members of the human population. One method of estimating these parameters is by using anthropometric measurements as inputs into a geometric humanoid model, in which each body segment is comprised of a certain number of regular volumetric shapes depending on the complexity of the shape and on the complexity of the model. Several of these geometric body segment parameter models have been proposed, among which the work of the late Herbert Hatze (1937–2002) stands out as the most detailed and accurate. Due to the mathematical complexity of his model (1979–1980), his work has not been widely used for biomechanical research; while often cited, it has rarely been reproduced. In this paper, an open source redevelopment of Hatze’s model is described, which includes a visualisation component that allows the 3D geometry to be accurately depicted for the first time. An analysis is performed which relates the accuracy of the model to the number of anthropometric measurements taken and suggests a reduced-order version of the model using interpolation. In order for the model to be more widely used and hopefully improved in time, the Matlab code for this work is made freely available for use by the research community.