Forums Anyscript.org  

Go Back   Forums Anyscript.org > General Forum > AnyBody Modeling System

Notices

Reply
 
Thread Tools Display Modes
  #1  
Old 01-12-2018, 09:25 PM
PhoebeHsu PhoebeHsu is offline
Senior Member
 
Join Date: Jul 2013
Posts: 100
Default How to average gait cycles?

Dear AnyBody:

I have a general question regarding using AnyBody to perform musculoskeletal modeling. Typically, the user choose the beginning and the end of a gait cycle, within which all variables (moments/forces) will be calculated. However, as shown in the attachment, there are always substantial cycle-to-cycle variances. Is it possible to let AnyBody run for a few cycles, and then output the average of kinematics/kinetics from those cycles?

thanks,
-Phoebe
Attached Images
File Type: jpg anybody.jpg (99.4 KB, 5 views)
Reply With Quote
  #2  
Old 01-15-2018, 10:37 AM
pgalibarov's Avatar
pgalibarov pgalibarov is offline
Senior Member
 
Join Date: Oct 2010
Posts: 960
Default

Dear Phoebe,

Not out of the box. The question is indeed tricky, because as you say yourself there is some variability even within cycles. You would need to extract start/end time, which may vary as well, and normalize all cycles to, say, 0-50-100 percent phase. This information is not intrinsic and needs to be computed. It is probably easier to do externally through Python/Matlab scripts.

It is also possible to treat these different cycles as new trials and just split all of them (using tStart/tEnd). Then when postprocessing the output - average by the subject/c3d file name.

Kind regards,
Pavel

Last edited by pgalibarov; 01-15-2018 at 10:41 AM.
Reply With Quote
  #3  
Old 01-22-2018, 09:58 PM
PhoebeHsu PhoebeHsu is offline
Senior Member
 
Join Date: Jul 2013
Posts: 100
Default

thanks Pavel. I shall investigate all the available approaches and report what I find to the community.

-Phoebe
Reply With Quote
  #4  
Old 01-25-2018, 10:31 PM
PhoebeHsu PhoebeHsu is offline
Senior Member
 
Join Date: Jul 2013
Posts: 100
Default

Quote:
Originally Posted by pgalibarov View Post
Dear Phoebe,

Not out of the box. The question is indeed tricky, because as you say yourself there is some variability even within cycles. You would need to extract start/end time, which may vary as well, and normalize all cycles to, say, 0-50-100 percent phase. This information is not intrinsic and needs to be computed. It is probably easier to do externally through Python/Matlab scripts.

It is also possible to treat these different cycles as new trials and just split all of them (using tStart/tEnd). Then when postprocessing the output - average by the subject/c3d file name.

Kind regards,
Pavel
Hi Pavel,

Can Anybody determine the start and end of each gait cycles timing and store it in the C3D file? Or those are defined by the user and AnyBody only does kinematic/kinetic calculation based on the selected frame(s)?

Thanks,
-Phoebe
Reply With Quote
  #5  
Old 01-30-2018, 11:50 AM
pgalibarov's Avatar
pgalibarov pgalibarov is offline
Senior Member
 
Join Date: Oct 2010
Posts: 960
Default

Hi Phoebe,

No, those should be defined by the user at this point.

Kind regards,
Pavel
Reply With Quote
  #6  
Old 03-07-2018, 11:17 PM
PhoebeHsu PhoebeHsu is offline
Senior Member
 
Join Date: Jul 2013
Posts: 100
Default

Quote:
Originally Posted by pgalibarov View Post
Hi Phoebe,

No, those should be defined by the user at this point.

Kind regards,
Pavel
Hi Pavel,

Can AnyBody read ascii files instead of c3d files, which include trajectory information of the physical markers and FP reaction forces?

Thanks,
-Phoebe
Reply With Quote
  #7  
Old 03-08-2018, 10:54 AM
pgalibarov's Avatar
pgalibarov pgalibarov is offline
Senior Member
 
Join Date: Oct 2010
Posts: 960
Default

Hi Phoebe,

Yes, of course. But you would need to implement your own soft marker drivers.

Pavel
Reply With Quote
  #8  
Old 03-08-2018, 06:10 PM
PhoebeHsu PhoebeHsu is offline
Senior Member
 
Join Date: Jul 2013
Posts: 100
Default

Quote:
Originally Posted by pgalibarov View Post
Hi Phoebe,

Yes, of course. But you would need to implement your own soft marker drivers.

Pavel
Hi Pavel,

Thank you for your reply!

As shown in the picture, I calculated the mean and standard deviation of the trajectories of each marker (the figure shows the Z- motion of the LKNE) during multiple cycles. Time was normalized to be between 0 to 100. GRF was processed in a similar way.

I haven't found a way to rewrite the data (ascii) to c3d so it can be directly read by AnyBody. If AnyBody can read kinematic information of the markers and the kinetic information of the FPs it will be great! Your thoughts are suggestions are greatly appreciated.

-Phoebe
Attached Images
File Type: png anybody.png (22.6 KB, 2 views)
Reply With Quote
  #9  
Old 03-08-2018, 06:20 PM
PhoebeHsu PhoebeHsu is offline
Senior Member
 
Join Date: Jul 2013
Posts: 100
Default

Quote:
Originally Posted by pgalibarov View Post
Hi Phoebe,

Yes, of course. But you would need to implement your own soft marker drivers.

Pavel
Not sure if I understand what you meant by "soft marker drivers".

Usually I will pick one cycle of this c3d file to run AnyBody. Now, I would like to use the average of multiple cycles of the same c3d files to drive AnyBody. I can get the kinematic info (i.e., markers' trajectories) and kinetic info (i.e., GRFs), both of which are in acsii format. If I didn't specify soft marker drivers when using normal c3d input, is it still necessary to specify them when using an acsii file derived from that same c3d file?

Thanks!
-Phoebe
Reply With Quote
  #10  
Old 03-15-2018, 10:19 AM
toerholm's Avatar
toerholm toerholm is offline
AnyBody Support
 
Join Date: Feb 2009
Posts: 927
Default

Hi Phoebe,

It is not a good idea to use average data as input to the analysis in AnyBody, for several reasons:
• For a segment with two markers on, in one trial their distance will be constant, after averaging of data this is no longer the case.
• After doing averaging on marker data the motion will have changed and there is no guarantee it will match the average of the GRF so you may see high residuals.


Best regards
Søren
__________________
TIP: How to save and load results.
http://wiki.anyscript.org/index.php/All_about_AnyBody_Modeling_System#Save_Load_and_Re play_Results

Subscribe to the forums, click the 'Forum Tools' link above the list of threads then select 'Subscribe to Forum'. You will then have the option to choose the subscription mode for this forum and get an email automatically.
Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +2. The time now is 11:39 AM.