Average velocity of CoP

[Petcu Daniel]

Dear Members,
I would like to analyze the velocity of the center of pressure as measured by an AMTI force platform for walking trials. I've obtained different values for the value of Average of CoP velocity between filtered (27.87 cm/sec) and unfiltered AMTI data (43.51 cm/sec) for the same original trial data. The average VCop given by AMTI-BioSoft analysis is 37.48 cm/sec for the same trial! I would like to understand which data should be used: filtered or unfiltered.
Any help will be appreciated !
Respectfully,
Daniel

[Peter Gabriel Adamczyk]

Re: Average velocity of CoP

Hi Daniel

COP in walking is highly sensitive to noise and to the magnitude of the vertical force.

If your velocity were a vector quantity, then the mean velocity would just be the starting and ending points of the COP, divided by the time it took to move between them. This measurement would be highly sensitive to the choice of Fz threshold, below which you consider the COP invalid (due to Fz --> 0, in the denominator of the COP derivation). But, it would not be sensitive to noise between these points, as rapid oscillations would cancel out.

On the other hand, using a scalar or "magnitude" when computing instantaneous Vcop (prior to averaging) will be highly sensitive to noise, as the rapid oscillations will increase the apparent velocity, rather than cancelling out. To reduce this noise effect, you'll want a filtered version of some kind, though the best cutoff frequency is unclear.

It seems you're using a scalar velocity, which will suffer from both the noise problem and the Fz problem. I would recommend using a filtered version of COP and a relatively high Fz threshold. The high threshold will cut off the ends of the COP, which may be undesirable. But since it's impossible to determine exactly when the COP transitions from "mostly correct" to "mostly artifact," you want to stay well away from that uncertainty. You might want to choose the cutoff frequency for your filter by sweeping across a range of choices, and picking something in the middle of a range where the outcome's sensitivity to cutoff frequency is minimal.

Good luck,
-Peter Adamczyk

[Petcu Daniel]

Re: Average velocity of CoP

Many thanks, Peter !
Respectfully,
Daniel