# Minimization of Metabolic Cost in Gait

It is known that humans, and other animals, choose a walking speed that minimizes metabolic cost. Therefore, an optimization that minimizes metabolic energy should improve predictions of gait. The goal is to implement a model of metabolic energy as an objective in our predictive simulation framework.

Several models exist that find metabolic energy as a function of muscle fiber length and velocity, and activation. First, these models are compared on existing predictive simulations of gait. They should be able to predict a relative change of around 1% for subjects when a small change is made to their dynamics.

Also, these models often have a discontinuity, meaning that the first and second derivative do not exist everywhere. These cannot be used in our direct collocation approach as an objective. Therefore, several adaptations should be made to make the model continuous and smooth. Then, predictive simulations of gait can be found that minimize metabolic energy.