Optimal control simulations of human arm motion

Abstract

Using multibody systems to represent bones and joints is a common way to simulate human motion. Often, this is done with inverse or forward dynamics (solving initial value problems). However, many simulation tasks in biomechanics lead to boundary value problems, like performing a motion from a given start to a prescribed end position, which could be performed in various ways. In this context, we suppose that human motion is controlled in order to perform an optimal motion. Hence we formulate an optimal control problem (OCP) and compare the results when using different physiologically motivated cost functions. A direct transcription method called DMOCC (see [1]) is used to solve the OCP, whereby we benefit from it's structure preserving formulation, as the resulting optimal discrete trajectories are symplectic-momentum preserving. (© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim)