Sensitivity of forelimb swing phase inverse dynamics to inertial parameter errors

Estimations of segmental inertial parameters are required for true inverse dynamics calculations during the swing phase of locomotion. This study attempts to quantify the effect of inertial parameter errors on inverse dynamic solutions. Swing phase forelimb net joint moments and powers at the trot (mean +/- s.d 3.03 +/- 0.16 m/s) were calculated with sagittal plane kinematic data from 5 Dutch Warmbloods using inertial parameters based on published regression equations for the breed. Significant peaks in the net moment and power curves for each forelimb joint were identified and measured. Net joint moments and powers were then recalculated after varying the segment mass, location of the segment centre of mass and the mass moment of inertia separately for each of the limb segments. Peak values for the net joint moments and joint powers were determined after each variation, and the percent change in peak value per percent change in inertial parameter was calculated. Segment mass was the most influential parameter, followed by location of the centre of mass. Changes in the mass moment of inertia showed little effect on peak values. The most influential single inertial parameter was the mass of the hoof segment with a mean +/- s.d effect 0.74 +/- 0.22 and 0.69 +/- 0.18 percent peak change per percent parameter change on net joint moments and powers, respectively, across all joints. The results demonstrate the need for an accurate approximation of segment masses during the swing phase, especially the hoof, and the need to account for any additional masses in the model, such as shoes.