The standard methods for human (and animal) movement research and clinical movement analysis involve motion capture systems which rely on markers. Often the markers are reflective balls. The marker-based systems measure the 3D trajectories of the markers. however, what the researcher and clinician is interested in is not the trajectories of the markers, but the movement of the joints and segments of the body. The problem of computing rigid body movement from marker trajectories is an inverse kinematics problem.

A common problem in motion capture is that the position of markers are missing during parts of the data set. The reason is often that the marker is occluded by body parts. We have developed a solution to this problem by formulating a state space model describing the kinematics of the body segments and joints, and tracking the state (motion) of the model with an Extended Kalman Filter. This makes it possible to update the filter with data from the currently available markers at each time step.

The model used should be as good an representation of the actual skeletal mechanism of the body as possible. One of the most important parts of the model is to get correct positions of the joints, which has inspired us to develop new and better methods for determining the position of the joint from measured movement of the body.

Publications

• Halvorsen K, Johnston C, Back W, Stokes V, Lanshammar H. Tracking the motion of hidden segments using kinematic constraints and Kalman filtering. J Biomechanical Engineering. 2008 Feb; 130(1).
• Halvorsen K, Söderström T, Stokes V, Lanshammar H. Using an extended Kalman filter for rigid body pose estimation. J Biomechanical Engineering. 2005 Jun; 127(3):475-83.
• Halvorsen K. Bias-compensated least squares estimate of the center of rotation. J Biomechanics. 2003 Jul; 36(7):999-1008.
• Halvorsen K, Lesser M, Lundberg A. A new method for estimating the axis of rotation and the center of rotation. J Biomechanics. 1999 Nov; 32(11):1221-7.