Adding GRF Prediction to GaitVaughan Model


#1

My current project is to simulate the differences of muscle activation between human body model with a knee exoskeleton on and without an exoskeleton during gait.

The AMMR model I chose is GaitVaughan model. I successfully connected the exoskeleton to human body model and ran inverse dynamic. But there was no difference between the body with exoskeleton and without it.

I thought the problem was Ground Reaction Force was not changed in GaitVaughan Model. Hence, I am solving this problem with replacing force plates to GRF Prediction algorithm which developed by AnyBody.

I added GRFPrediction to GaitVaughan model as Squat AMMR model did. But I got some errors to load model, errors are shown as:

  1. When I use CreateFootPrintClass.any and CreateFootConnectionClasses.any, I got error:

ERROR(SCR.PRS9) : C:/U…s/l…i/D…p/a…y/G…n/M…l/FootPrintGeneration.any(8) : Defined at : C:/P…s/A…y/A…1/AMMR/T…s/M…s/F…s/CreateFootPrintClass.any(32) : ‘RefSource3’ : Unresolved object
2. When I do not use CreateFoorPrintClass Modelutils, I got error:
ERROR(SCR.PRS9) : C:/U…s/l…i/D…p/a…y/G…n/M…l/GRFSupports.any(13) : Defined at : C:/U…s/l…i/D…p/a…y/G…n/G…n/CreateFootNodes.any(14) : ‘HeelContactNodeLow’ : Unresolved object

​Could you please tell me how to add GRFPrediction to other models instead of ANYMocap models?
How to use CreateFootPrintClass.any and CreateFootConnectionClasses.any correctly? Or how to define the fours nodes which are needed to create foot nodes?

I also downloaded GRF_Prediction_master model from WiKi, but it can not be loaded with error:
ERROR(SCR.SCN6) : C:/P…s/A…y/A…1/AMMR/Body/A…n/B…s/G…l/BodyModel.any(16) : ‘BM_LEG_RIGHT’ : Unexpected character.


#2

Hi Xiying,

Welcome to the forum.

I will try to answer your questions.

First of all the GaitVaughan is not a good model to start out with in my opinion, it based on an old leg model and an old modeling approach originating from that the data is not placed in a C3D file like in the more recent models.

For these reasons i would strongly recommend to shift the model to be one the model in the "MocapExamples " folder. This will be easier and give you better results.

The reason why see no difference can be the following:

[ol]
[li]The exo model has not been include into the study in the correct way, so the model is being loaded but is not part of the study. This may happen if the exo files are not being include from within the study or if the exo model is not being refered to from the study.
[/li][li]If the exo has no mass and if there are no passive forces applied it may also lead to no difference in force.
[/li][/ol]

Concerning the GRF prediction it will be much easier for you to start out with model already setup with this like eg. this model “MocapExamples\Plug-in-gait_Simple\FullBody_GRFPrediction.main.any”

The reason why you get the error you see is that the old leg model is missing certain nodes to be defined on the foot segment. Additionally the GaitVaughan model do not contain fullbody mocap which is needed when working with GRF predictions.

The last error you are mentioning must be due to a mismatch between the GRF model you have downloaded and the AMMR model. The GRF model on the wiki is depricated as it is noted on the page
“Note: This guide is deprecated since Ground Reaction Prediction is now part of the AMMR. Please see the MoCap GRF prediction example.”

So please use the model i listed above it is an updated version of the model on the wiki.

Please include your affiliation in your personal info page.

Best regards
Søren