Automated Fitting of an Elastokinematic Surrogate Mechanism for Forearm Motion from MRI Measurements J. Xu 1 , P. Kasten 2 , A. Weinberg 3 and A. Kecskem´ ethy 1 1 Chair for Mechanics and Robotics, University of Duisburg-Essen, Germany; e-mail: {jing.xu, andres.kecskemethy}@uni-due.de 2 Department of Orthopaedic Surgery, University of Dresden, Germany; e-mail: Philip.Kasten@uniklinikum-dresden.de 3 Department for Pediatric Surgery, University of Graz, Austria; e-mail: annelie.weinberg@t-online.de Abstract. Forearm rotation (pro-supination) involves a non-trivial combination of rotation and translation of two bones, namely, radius and ulna, relatively to each other. Early works regarded this relative motion as a rotation about a fixed (skew) axis. However, this assumption turns out not to be exact. This paper regards a spatial-loop surrogate mechanism involving two degrees of freedom with an elastic coupling for better forearm motion prediction. The model parameters are not measured directly from the anatomical components, but are fitted by reducing the errors between predicted and measured values in an optimization loop. For non-invasive measurement of bone position, magnetic resonance imaging (MRI) imaging is employed. We present a method to self-calibrate the arm position in the MRI scanning tube and fitting the model parameters from a few, coarse MRI scans. Results show a good concordance between measurement and simulation. Key words: Forearm kinematics, surrogate mechanism, MRI motion fitting. 1 Introduction Performance of forearm pro- and supination motion (i.e., forearm rotation from palm down to palm up and vice versa), is one of the crucial functional motions of the upper extremity, which is used in everyday tasks such as turning a door handle or fixing a screw with a screw-driver. This motion is of particular interest in practice due to an increasing number of surgery treatments of the forearm. According to stat- istical data, for children, 20% of all recorded fractures occur in the forearm, making this the most frequent form of extremity fractures [1]. Therefore, there is a need for a clear description of forearm motion, especially under variable physiologic loading conditions and at different rotational positions of the forearm. Forearm motion research began in the early 20th century. Fick in 1904 presented the first kinematic model describing forearm motion, which stated that the forearm rotated around the ulna around a constant axis of rotation [2] (see Fig. 1a). How- ever, such a simplified model renders an unrealistic strong tilting in the maximal pronation. Based on MRI findings, more comprehensive kinematic models were de- veloped that include the motion of the ulna. Kapandji and Nakamura et al. stated 349 Machine, DOI 10.1007/978-90-481-9262-5_37, © Springer Science+Business Media B.V. 2010 J. Lenarčič and M.M. Stanišić (eds.), Advances in Robot Kinematics: Motion in Man and