AMTI force platform coordinate, 2D rotation matrix

[Zheng Wang]

Re: AMTI force platform coordinate, 2D rotation matrix

Hi Dr. Seeholzer and Dr. Chua,
Thank you for your comments and valuable suggestions. The problem has been solved. We called Vicon yesterday in the morning. We checked the A/D calibration matrix at the beginning. There was nothing wrong in there. We then noticed that we assigned a wrong A/D pin channel on the Nexus software. Our force platform has 8 channels. We somehow assigned the 10th channel of the 2nd platform as channel 18. By modifying the system file (re-assign the pin channel back to 10), we can retrieve the correct data without any problem :-)

[Thomas Seeholzer]

Re: AMTI force platform coordinate, 2D rotation matrix

Dear Zheng Wang

... one more thing:
Have you tried using AMTI NetForce to acquire the forces and COP data?
This will eliminate all further processing (spacial coordinate transformations) done by Vicon or any other additional software.

Regards,

Thomas.

[Matthew Chua]

Re: AMTI force platform coordinate, 2D rotation matrix

Is it possible that the wrong calibration matrix file has been applied to the force plates?

[Thomas Seeholzer]

Re: AMTI force platform coordinate, 2D rotation matrix

Dear Zheng Wang

My first guess is that you have some kind of crosstalk between the force plate's channels, i.e. the signals are mixing up before being converted to digital values.
Have you checked the wires?

Regards,

Thomas.

[Zheng Wang]

Hi,
In our study, we plan to detect the postural limitation boundary in a side-by-side foot position for each participant by requesting them leaning forward, backward, to the left and to the right with their ankle and hip joints' motion. The COP pattern should look like a cross with the maximum values (postural limit) at each direction. However, the COP data we have been collected from one force platform looks very odd (we have 2 force platforms, the 1st one looks reasonable, the 2nd one looks odd). The whole COP pattern deviated from the original coordinate. For example, if the participant leans forward, the resultant GRF (also, the COP) moves to the up-left corner of the platform. (Sorry, I couldn't upload my figure).

At first, I think the coordinate of the force platform is not aligned along the x and y directions (We've emailed Vicon on this matter). In order to fix the data though, I run PCA on my COP data to figure out the deviation angle and then applied a 2D rotation matrix to rotate the COP back to its original coordinate.

However, the COP traces still shows deviation from one axis. Then, I realized the x and y axes might not orthogonal to each other!!! If this is the case, I don't think 2D rotation cannot help me solve the whole problem. My feeling for my data is that 1) the COP x and y axes are not orthogonal to each other; and 2) each of axis has deviation from its own x and y directions.

If this is the case, should I trash my data? or there is any other way to save my data. Any comments and suggestion are appreciate.