Teach-in '98 on GRV & CoM: What people said...

Hi all,

In Posturography we see an increasing tendency that people want to look
at movements of the CoM in addition to the CoP measured with a force
plate.
Calculating CoM seems to be a simple double integration of the shear
forces of the force plate. However this doesn't work in practice because
the double integration enlarges the smallest errors in your measuring
chain - resulting in a CoM miles away of your starting point.
Even when you force your starting and your ending point to be the same
(by playing with the initial conditions) your CoM path is not
reasonable.

I have recently thought up a solution for the above problem:
*       At the instant when the shear force is zero, the CoM and CoP are
exactly above each other.
*       This would allow the design of a double integration algorithm
which only has to integrate the short periods of time between these
instants.

Has anybody come across this method before? And if yes, is it verified?

Best Regards
Christian Calame

Mr. Christian Calame, Product Manager Biomechanics
Kistler Instrumente AG Winterthur, P.O.Box 304,
CH-8408 Winterthur, Switzerland
Tel: +41 52 224 11 11, Fax: +41 52 224 14 14
http://www.kistler.ch/biomech



I and my supervisor have been discussing the Com and CoP problem. It
would indeed be valuable to have an accurate method for estimation of
the position of CoM from force plate data. I believe Mr. Calame's
proposed method to estimate CoM with a double integration of the shear
forces between timepoints when the shear force is zero, will work fine
if the body behaves exactly like an inverted pendulum. In this case the
ground reactin force vector, starting at the CoP, is pointing at the
CoM, and the horizontal component is zero only when the CoM projection
coincides with the CoP.

However, in reality we are not perfect inverted pendulums and I think
there are situations where the shear force can actually be zero but the
projection of CoM and CoP do not neccesserily have to coincide. I hope
this example will make my train of thought easier to understand:

Imagine a man with a wooden leg, where his feet have only a point
contact with the floor. This man can produce a shear force, for instance
by using hip flexors, which will cause his projection of CoM to move
backwards and away from the point support surface. To try to prevent
falling he then produces a shear force in the opposite direction by
using hip extensors. Sometime between these two actions the shear force
must be zero, but the CoP (which can not move from the point support
surface) and CoM will not be exactly above each other, since CoM has
moved a little backwards.

I would also like to comment the idea of a normative data base. I think
the idea is splendid, and I hope the data base will include postural
measurements as well. I definitely would like to contribute to such a
data base. By the way who will have the copyright of the data?

Best regards
Annica Karlsson

Annica Karlsson, Ph.D student
Systems and Control, Uppsala University
PO Box 27
751 03  Uppsala, Sweden
Tel: +46 18 741 78 46
E-mail: Annica.Karlsson@SysCon.uu.se,
http://frej.teknikum.uu.se/Personnel/ack/ack.html


Back to Teach-in