Recursive Filtering of Kinetic and Kinematic Data for Center of Mass and Angular Momentum Derivative Estimation

Estimating the center of mass position and the angular momentum derivative of the human body is an important topic in biomechanics, since both quantities are essential to the dynamic description of the motion. In this work, we introduce a novel recursive algorithm to accurately estimate them, by fusing kinetic and kinematic measurements, based on a spectral description of the noise carried by each signal. This method exploits the mathematical relationships that links the center of mass position and the angular momentum derivative to recursively improve their estimation. The effectiveness of the approach is demonstrated on a simulated humanoid avatar, where access to ground truth data is granted. The results show that our method reduces the estimation error on the center of mass position with regard to kinematic estimation alone, in addition to providing a good estimate of the angular momentum variation. The proposed framework is finally applied to a recorded human walking motion in order to illustrate its applicability to real motion analysis data.