Controlling Biomechanical Models To Move Like Humans Do

http://www.stanford.edu/~smenon/professional_files/publications/MenonKhatib-BMECS11.pdf

The above paper, while brief, cites a Stanford University study which attempts to quantify the ability of computer models to simulate human motion.  In particular, the researchers attached sensors to key body areas of a normal person, and observed the result.  In particular, the researchers noticed that the computer model constructs did not perfectly conform to the human body.  The researchers then used a novel idea: use vector transformation matrices in order to accurately account for small deviations from what was expected.  The result?  A much more accurate model of human biomechanical motion.

Why is this important?

As noted by the authors, an accurate understanding of the human gait, and the ability to apply that knowledge to other facets of human motion, is key in many medical disciplines, and can be used to predict rehabilitation strategies, common injuries associated with specific tasks, and other pathological and analytic trends.