Identifying Musculo-tendon Parameters of Human Body Based on the Musculo-skeletal Dynamics Computation and Hill-Stroeve Muscle Model Gentiane Venture, Katsu Yamane and Yoshihiko Nakamura Department of Mechano-Informatics University of Tokyo 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan gentiane@ynl.t.u-tokyo.ac.jp, yamane@ynl.t.u-tokyo.ac.jp, nakamura@ynl.t.u-tokyo.ac.jp Abstract— The goal of the works presented in this paper is to study the feasibility of the in-vivo estimation of the musculo- tendon dynamics using an original approach. Contrarily to usual literature in this field the Hill-Stroeve based model parameters are estimated, and not only some global parameters with little physiological meaning. This estimation is based on the computation of the inverse kinematics and the inverse dynamics of the human body, using a musculo-skeletal description. It also includes an optimization of the muscle force. Muscles are described by a modified and simplified Hill-type model. Parameters involved in this model are estimated in-vivo. In the preliminary works the flexion-extension of the elbow joint only is considered. Experimental set-up includes EMGs acquisition and motion capture data. Movements are chosen to limit co- contraction of antagonist muscles. Results are given for three of four muscles involved in the joint movements. Such results are very important for medical applications in rehabilitation, sport science, study of muscle diseases... I. I NTRODUCTION Robots get closer to human due to the miniaturization of components and the developments of artificial organs such as muscles and the enhanced computation power. Although the human body is very complex and the ability of movement is very wide which make its comprehension more difficult. For a clear understanding of the human dynamics and to generate smooth movements it is very important to have a good knowl- edge of the human anatomy and of its dynamics. Though actuators of the human body are the muscles, their dynamics is very important in the study of the human body movements. Its complexity is mainly due to its biologic nature, the high number of muscles and degrees of freedom and also because of the differences between each human being: size, mass, strength, capacity... which are closely related to the history of the subject: doing sports, having had injuries... In dynamic modelling it is important that the model used gives a realistic description of the behavior and that this model is adapted to the subject. For this reason the subject specific musculo- tendon’s dynamics must be estimated. The model used for the muscle modelling is based on the Hill-Stroeve model, known to be the most accurate model to describe the complex muscle CE l mt l m l t muscle tendon Fig. 1. Musculo-tendon complex without pinnation angle dynamics, and related to the micro-structure of the muscle. Unlike usual works on the human body where this model is globally scaled to the subject, the proposed method focuses on the subject-specific parameters involved in this model. Previous simulation results have shown that it is technically possible to estimate those dynamic parameters of the muscles. This paper first presents a sum up of the dynamic modelling of musculo-tendon complex based on the Hill-type modelling (Section II). Then in the third section a simple modelling of the elbow joint, considering one degree of freedom, is given. The fourth section presents the experimental set up based on the use of EMGs measurements and motion capture system. Finally the fifth section gives the identification method, some simulation results and the preliminary experimental results. II. MUSCULO- TENDON COMPLEX DYNAMICS The musculo-tendon complex is composed of the tendon and the muscle (Fig.1). The tendon is a passive wire that does not generate movement. The muscle is an active contractile element (generally noted CE) that generates contraction of the muscle controlled by neural excitation and then its lengthen- ing or its shortening. Contractions of the muscle are assumed to be iso-volume [1]. A. Musculo-tendon complex dynamics modelling The musculo-tendon dynamics is function of the muscle activity a(t), and the length and velocity of the muscle and the tendon, respectively l m , ˙ l m , l t and ˙ l t . The model given here Proceedings of 2005 5th IEEE-RAS International Conference on Humanoid Robots 0-7803-9320-1/05/$20.00 ©2005 IEEE 351