DEALING WITH SKIN MOTION AND WOBBLING MASSES IN INVERSE DYNAMICS

Abstract

Inverse dynamics is a standard analysis in biomechanics to reconstruct time histories of internal driving forces and torques from measured external forces and segmental kinematics. The main sources of inconsistency leading to analytical artefacts in this process are skin marker and soft tissue motion. These potentially artificial high frequency fluctuations in the joint torques may serve as an erroneous basis of (misleading) assumptions with respect to muscular activity. Here we suggest techniques to reduce these errors. In both parts of this study, high-speed video and force platform data were acquired. In one part, 69 sequences of human barefoot running were sampled followed by an inverse dynamic analysis of the stance leg. The time history of the hip joint torque in the sagittal plane served as a sensitive "detector" of dynamic analysis artefacts. We show that the most important error — the relative skin to bone motion especially of the knee marker — can be reduced significantly by processing kinematic data using bone rigidity (constant segment lengths) and bony contour (frontal knee edge) information. Further on, neglecting significantly initiated soft tissue dynamics in the inverse dynamic model introduces another inconsistency in the analytical process. Therefore, in a second part of this study, soft tissue kinematics from 14 jumping sequences were identified. These data provided a set of coupling parameters of wobbling masses to the bone that were ready to be implemented in the inverse dynamic model. Using realistic bone kinematics mainly avoids phase shifts in the acceleration scenario within the leg, and thus artifical hip torque fluctuations within the whole contact period. In human running, accounting for soft tissue dynamics mainly affects the calculated timing of the hip joint torque during the impact phase.