Optimization-based dynamic human walking prediction

Abstract

In this study, an optimization-based approach for simulating the walking motion of a digital human model is presented. A spatial skeletal model having 55 degrees of freedom is used to demonstrate the approach. Design variables are the joint angle profiles. Walking motion is generated by minimizing the mechanical energy subjected to basic physical and kinematical constraints. A formulation for symmetric and periodic normal walking is developed and results are presented. Backpack and ground reaction forces are taken into account in the current formulation, and the effects of the backpack on normal walking are discussed. Copyright © 2007 SAE International.