An Approach Experimentally Based Modeling and Simulation of Human Motion

A computational approach for estimating three-dimensional human body motion from measurement of marker locations during motion is proposed. The proposed method computes the system’s kinematics while preserving important physical and dynamic properties. These properties include preserving the connection point (joint) between any two neighboring bodies and satisfying total body linear and angular momentum conservation laws. Four sets of equations are formulated for kinematic and dynamic properties of body segments. Attitude estimation is based on Wahba’s problem [1], and dynamic and kinematic constraints are included by utilizing penalty function method. An iteration process is presented to combine the four sets of estimation measure to ensure convergence to the optimal solution. Two examples are presented to demonstrate the performance of the proposed method: estimation of the kinematics of a chain of rigid bodies obtained by computer simulation, and estimation of motion in three dimension of a diver obtained using experimental measurements. The results of both examples show fast convergence of the algorithm to an optimal solution while satisfying the imposed the constraints.