OpenSim Webinar - Metabolic cost modeling: experimental validation and predictive simulations

The OpenSim project and the National Center for Simulation in Rehabilitation Research (NCSRR) at Stanford invite you to join our next webinar, Anne Koelewijn, D.Eng. from Friedrich-Alexander Universität Erlangen-Nürnberg & Antonie van den Bogert, PhD from Cleveland State University. 

DETAILS 
Title: Metabolic cost modeling: experimental validation and predictive simulations
Speakers: Anne Koelewijn, D.Eng., Friedrich-Alexander Universität Erlangen-Nürnberg & Antonie van den Bogert, PhD, Cleveland State University
Time: Thursday, September 26, 2019 at 10:00 a.m. Pacific Time
Register: There are limited spots. Please register ahead of time.

ABSTRACT 
It is not known whether muscle activation or metabolic energy is minimized during human movement. In order to investigate this question with musculoskeletal models, a mathematical model for metabolic energy expenditure is needed. We investigated the validity of seven metabolic energy models by comparing their estimates to oxygen uptake data. The models by Bhargava et al. (J Biomech, 2004: 81-88) and Lichtwark and Wilson (J Exp Biol, 2005: 2831-2843) had the best agreement with the oxygen data. Next, we compared predictive gait simulations that minimized metabolic cost against a simulation that minimized squared muscle activation. Joint angles were more realistic with metabolic cost minimization, while joint moments were more realistic when minimizing activation.
 
In this webinar, we will discuss the experiment and analysis used to calculate metabolic cost from gait data and compare the metabolic energy models. Then, we will explain how the metabolic energy models were modified to be twice differentiable, which is required for gradient-based optimization. Finally, we will explain how we found predictive gait simulations using objectives of metabolic cost and squared muscle activation and how we compared these objectives.
Webinar Start Date
Webinar End Date