Anybody Model - Walking


#1

Hello,

Just a question!

Are there standard or well established musculoskeletal application model for Walking?
It is much appreciated if some of you can give me a pointer on this.

Thank you.


#2

Hello,

First of all, could you please provide your affiliation and full name for us to know whom we are supporting? Many thanks.

Secondly, we do not have a “standard” walking application. But we have an application for processing motion files recorded through by different Motion Capture systems. This is as standard as it gets.

You can find it in:
ammr\Application\MocapExamples\

Kind regards,
Pavel


#3

Hello,
Thank you for your response. I did add my affiliation and details.

Just have another question.

I am currently working on “Lower Extremity walking application”. There are few gait data files (C3D files) in input folder. Rather than using them, can we work using another gait data file?
How can we convert normative gait data (excel files) into C3D files?

Thank you.


#4

Hi AHW,

The mocap models also runs with other C3D data, it is relatively simple to modify them to use other C3D files.

Honestly i am not sure how to do that in a good way the models are setup to use a C3D file as input.

The model run in two steps the first step extracts the kinematics from the model and saves this as joint angles and pelvis motion. In principle if you could translate the normative gait data to this format it may work, but note that joint angles will not be enough you will need also the rigid body motion of e.g. the pelvis to fully describe the kinematics. You can try to run the current model and look into the output data it generates for the kinematics and see if you have such data. Secondly you will to build into the model the same kinematics measures that was used to generate the normative data, so all joint angle definitions needs to be same, this is in principle not a problem you can define new kinematic measures in your model that corresponds to the data you have.

So it might be possible it depends on the data you have, but it will require some work. Have you considered to use C3D data from eg http://mocap.cs.cmu.edu/ here you will find motion data but no GRF, but these might be predicted.

Best regards
Søren

Best regards
Søren


#5

Hello,

Thank you for your response.

I have already tried it and got lot of errors in force plates. With the help of other forums I have found solutions for them. But now I am stuck at kinematic analysis. It said “Kinematic analysis is failed in step 0, check initial conditions”. I did lot of modifications on any script and look. It didn’t fix yet.

Can anyone give me a suggestion to update initial conditions?


#6

Hi AHW,

If the model fails to solve the first timestep it can be a sign that the initial conditions are not good enough (close enough to the solution of the first frame), but it can also mean that drivers you have on the model are not “correct”, or it can be both.

I would always recommend to take a few steps backwards in such a situation, so try to simply the model to make it run… so e.g. remove the arms, add constant drivers in specific joints etc.

See this wiki post on debugging kinematics. https://github.com/AnyBody/support/wiki/Model-trouble-shooting#resolve-kinematic-problems

Hope it helps
Best regards
Søren


#7

Hello Søren,

I did tried several methods to fix it. Finally rather than using my own C3D file, modified given Lower extremity input C3D file (Gait Normal_1). I removed upper extremity markers as similar to my own C3D file. That’s the only modification I did. Then load the model and run analysis. Now I got the following error.

[i]0) Operation Sequence: (Operation: Main.RunAnalysis):
0.0) Operation Sequence: (Operation: Main.RunAnalysis.LoadParameters):
0.0.0) Dummy operation: (Operation: Main.RunAnalysis.LoadParameters.placeholder_operation):
0.1.0) Load_parameters (Operation: Main.ModelSetup.Macros.Load_parameters):

Macro command : Load_parameters(1:10)> classoperation Main “Load Values” --file=“C:\Users\DELL\Documents\Anybody\AnyBody.7.1.x - modified model\AMMR.v2.1.0-Demo\Application\MocapExamples\Plug-in-gait_Simple\Output/LowerExtremity_GaitNormal_3.anyset”

Loading and overriding values from : ‘C:\Users\DELL\Documents\Anybody\AnyBody.7.1.x - modified model\AMMR.v2.1.0-Demo\Application\MocapExamples\Plug-in-gait_Simple\Output\LowerExtremity_GaitNormal_3.anyset’ …
ERROR(SCR.SCN1) : Cannot open file : C:\Users\DELL\Documents\Anybody\AnyBody.7.1.x - modified model\AMMR.v2.1.0-Demo\Application\MocapExamples\Plug-in-gait_Simple\Output\LowerExtremity_GaitNormal_3.anyset

Macro command : Load_parameters(10:10)> classoperation Main “Update Values”

Updating expressions…
…Finished updating expressions
1.0) Operation Sequence: (Operation: Main.RunAnalysis.MarkerTracking):
1.0.0) Dummy operation: (Operation: Main.RunAnalysis.MarkerTracking.placeholder_operation):
1.1.0) Kinematics (Operation: Main.Studies.MarkerTracking.Kinematics):
1.1.0.0) PreOperation (Operation: Main.Studies.MarkerTracking.Kinematics.PreOperation):
1.1.0.0.0) InitialConditions (Operation: Main.Studies.MarkerTracking.InitialConditions):
1.1.0.0.0) …Design variables have been updated.
1.1.0.0.1) …Load-time positions have been re-established.
1.1.0.0.1.0) PreOperation (Operation: Main.Studies.MarkerTracking.InitialConditions.PreOperation):
1.1.0.0.1.0.0) KinematicView (Operation: Main.ModelSetup.Views.SetViewMacros.KinematicView):

Macro command : KinematicView(1:1)> classoperation Main.ModelSetup.Views.KinematicView"Set View"

ERROR(OBJ.MCH.KIN3) : C:/U…s/DELL/D…s/A…y/A…l/A…o/T…s/A…p/AnyMoCapModel.any(41) : MarkerTracking.InitialConditions : Kinematic analysis failed in time step 0 : Indefinite positions[/i][/COLOR]

It seems to be an error in marker tracking initial conditions. Can you suggest me a way or guide me t solve this problem.

Thank you.
Amashi


#8

Hi Amashi,

The error says indefinite positions, this normally means that one or more dof in the model is not driven.

From your description i understand you removed markers on the upperbody, if this is the case and the arms are still in the model they need to be driven by something else than markers otherwise the model is undetermined.

You can add soft drivers for these joints see e.g. https://anyscript.org/ammr-doc/bm_config/mannequin.html

This approach assumes you are using a recent version of AMMR otherwise the BM statements may have no effect, if this is the case the drivers will need to be added manually.

Bets regards
Søren