In response to Richard Smith's question about measuring the angular
orientation of the upper arm from two markers on the upper arm and one
on the forearm, may I suggest the following method:

I do not use markers, rather I digitize the shoulder joint center, the elbow
joint center, and the wrist joint center (in 3-D) from film or video. The
shoulder flexion/extension angle and ab/adduction angle can be computed from the
shoulder and elbow points relative to points on the trunk. To get
internal/external rotation, however, you need the wrist point as well. As
long as the elbow does not approach full extension, you can compute the cross
product of the upper arm vector and the forearm vector to get the elbow joint
flexion axis vector. The orientation of this axis in the transverse plane
of the humerus defines its internal/external rotation angle. If you need to
use reflective markers for an automatic digitizing system, however, you may
introduce errors by not locating the true joint centers with this method.

This elbow flexion axis is not defined when the forearm and upper arm are
parallel because of the limitations of the cross product. In reality, because
of the "carrying angle" of the forearm relative to the upper arm, the cross
product will produce an elbow flexion axis pointing in the AP direction when
the elbow is fully extended. I have not figured out a solution to this
problem.

I hope this was helpful. Other people's comments here would be appreciated.

Sincerely,

Rick Hinrichs
Dept. of Exercise Science
Arizona State University
Tempe, AZ 85287-0404 USA