This paper addresses the problem of estimating human body dynamics from 3-D visual data. That is, our goal is to estimate the state of the system, joint angle trajectories and velocities, and the control required to produce the observed motion from indirect noisy measurements of the joint angles. For a two-link chain in the human body, we show how two independent spherical pendulums can be composed to create a behaviorally equivalent double spherical pendulum. Therefore, the estimation problem can be solved in parallel for the low-dimensional spherical pendulum systems and the composition result can be used to arrive at estimates for the higher dimensional double spherical pendulum system. We demonstrate our methods on motion capture data of human arm motion.
[1]
Pietro Perona,et al.
Decomposition of human motion into dynamics-based primitives with application to drawing tasks
,
2003,
Autom..
[2]
Richard M. Murray,et al.
A Mathematical Introduction to Robotic Manipulation
,
1994
.
[3]
M. Pitt,et al.
Filtering via Simulation: Auxiliary Particle Filters
,
1999
.
[4]
Christoph Bregler,et al.
Learning and recognizing human dynamics in video sequences
,
1997,
Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition.
[5]
Victor M. Becerra,et al.
Optimal control
,
2008,
Scholarpedia.