Optimization and Validation of a Biomechanical Model for Analyzing Running-Specific Prostheses

摘要

Modeling the ankle joint during amputee locomotion is difficult since a definitive joint axis may not exist. Gait analysis estimates joint center positions and defines body segment motions by placing reflective markers on anatomical landmarks. Inverse dynamics techniques then estimate joint kinetics (forces and moments) and mechanical energy expenditure using data from ground reaction forces (GRFs) and the most distal joint (usually the ankle) to make calculations for proximal joints. Running-specific prostheses (RSPs) resemble a "C" or "L" shape rather than the human foot. This allows RSPs to flex and return more propulsive energy, like a spring, but no "ankle" exists. Current biomechanical models assume such a joint exists by placing markers arbitrarily on the RSP (e.g. the most acute point on the prosthesis curvature). These models are not validated and may produce large errors since inverse dynamics assumes rigid segments between markers but RSPs are designed to flex. Moreover, small errors in distal joint kinetics calculations will propagate up the chain and inflate errors at proximal joints. This study develops and validates a model for gait analysis with RSPs. Reflective markers were placed 1 cm apart along the lateral aspects of five different RSPs. Prostheses were aligned in a material testing system between two load cells. Forces simulating peak running loads were applied and the load cells measured forces and moments at the top (applied force) and bottom (GRF) of the prostheses. Inverse dynamics estimated force transfers from the bottom to top of the prostheses through the defined segments. Differences between estimated and applied values at the top are considered model error. Error will be calculated for every possible combination of markers to determine the minimal marker set with an "acceptable" level of error. The results yield a model that can be confidently used during gait analyses with RSPs.