Page 1 of 1

tracking global coordinates of a point......

Posted: Tue Aug 25, 2009 9:54 am
by ananthgk
Hi

Is it possible to track the 'global' coordinates (in the lab frame), of a certain point like the origin/insertion point of a muscle, center of mass of a segment etc. for each time-step of a walking trial? I can get the data in the local frames from the XML files, but i want to track the movement of these points during a walking trial and export the data to another program in MATLAB, so it would help if there was a way to obtain this data in a .trc file or something similar. Thanks a lot!!

Ananth

RE: tracking global coordinates of a point......

Posted: Fri Aug 28, 2009 1:56 am
by aymanh
Hi Ananth,

It is possible indeed to track the global coordinates of a point using CMC. You have to use the task type rdCMC_Point (in the CMC tasks file). Please check the documentation of this Class for details. The point can be specified in the frame of any Body.

Hope this helps,
-Ayman

RE: tracking global coordinates of a point......

Posted: Mon Aug 31, 2009 8:23 am
by cucho
Hi,

I have been trying to track muscle insertion points. I have seen the documentation in the user guide and have changed the rdCMC_Joint to rdCMC_Point. Then set the insertion point coordinate under the </coordinate> headings. The program crashes everytime, I am not shure what I am doing wrong.

Thanks,

Daniel

RE: tracking global coordinates of a point......

Posted: Mon Aug 31, 2009 10:24 am
by ambaus
So, I have done this exact thing. I've written Matlab code to extract BodyKinematics analysis data and transform muscle points into the global frame.

RE: tracking global coordinates of a point......

Posted: Mon Aug 31, 2009 2:15 pm
by aymanh
Daniel & Anne,

A coordinate is supposed to refer to an actual coordinate in the model!

Following is the XML representation for an rdCMC_Point task. You can obtain the same info by running the command:
>cmc -PI rdCMC_Point

Hope this helps,
-Ayman



<rdCMC_Point name="default">
<on> true </on>
<!--Name of body frame with respect to which a tracking objective is specified. This property is not used for tracking joint angles.-->
<wrt_body> </wrt_body>
<!--Name of body frame in which the tracking objectives are expressed. not used for tracking joint angles.-->
<express_body> </express_body>
<!--Array of 3 flags (each true or false) specifying whether a componentof a task is active. For example, tracking the trajectory of a point in space could have three components (x,y,z).A task for tracking a joint coordinate only has one component.-->
<active> false false false </active>
<weight> 1.0000 1.000 1.000 </weight>
<kp> 1.0000 1.0000 1.0000 </kp>
<kv> 0.5000 0.50000 0.5000 </kv>
<ka> 1.0000 1.00000 1.000</ka>
<!--Direction vector[3] for component 0 of a task. Joint tasks do not use this propery.-->
<r0> 0.0000 0.0000 0.000 </r0>
<r1> 0.0000 0.0000 0.000 </r1>
<r2> 0.0000 0.0000 0.000 </r2>
<point> 0.0000 0.0000 0.000 </point>
</rdCMC_Point>

RE: tracking global coordinates of a point......

Posted: Mon Aug 31, 2009 5:37 pm
by aseth
There is a PointKinematics analysis that reports the location of a point in ground frame. Use the AnalyzeTool to add a PointKinematics analysis (one per point of interest) and edit it to specify the body that the point is on and its coordinates in that body. A file is generated with the name of the point in filename.


RE: tracking global coordinates of a point......

Posted: Mon Aug 31, 2009 5:46 pm
by ambaus
Ahh..that is what I need. However, when tracking the line of action of muscles, sometimes via points turn on and off. My code looks at the general coordinates and calculate whether the via point is used or not.

RE: tracking global coordinates of a point......

Posted: Tue Sep 01, 2009 9:19 am
by aymanh
Hi All,

Sorry about the misunderstanding on my part. I assumed you're solving the control problem of trying to find excitations that causes a point on the model to "follow a specified trajectory".

If all what you want is reporting the coordinates of the point then please ignore this, and use the PointKinematics analysis as Ajay suggested on this thread. Thanks Ajay!

-Ayman

RE: tracking global coordinates of a point......

Posted: Mon May 17, 2010 3:41 pm
by ambaus
I can't get PointKinematics to work. It always crashes.