A method to evaluate human skeletal models using marker residuals and global optimization

Different models are proposed in the literature to study the musculo-skeletal system. These models differ in several aspects like the number of segments in which the system is divided or the way to model articular joints. Marker residual analysis is a widely used method to evaluate how well the model fits the experimental measurements. However, the same model can yield different values of marker residuals depending on the PGA (Protocol of Gait Analysis) employed. The goal is to analyse the changes in the marker residuals obtained with different PGAs when a certain mechanical parameter is modified. Two of the most important parameters in the kinematic analysis have been studied: the length of a segment and the mechanical model of a joint. The results show that marker residuals obtained using some PGA do not provide valuable information to decide the best value of the analysed parameter. Values of the residuals computed using a PGA where the position of the markers is treated with the GOM (Global Optimization Method) are proposed to quantify the effect of certain parameters in the model.