Modelling knock knees or bowlegs?

Main Forum
1

Hi,

I am new to AnyBody system and I would like to run a gait analysis of a model with knee misalignment. Is there any way I can rotate the knee joint inward/outward to represent knock knees/bowlegs? Which leg models allow this adjustment?

Thanks in advance!

2

Hi,

If you see the structure of our human model then you can find that these two AnyRefNode objects are used for the definition of the knee joint:
Main.HumanModel.BodyModel.Right.Leg.Seg.Thigh.KneeJoint
Main.HumanModel.BodyModel.Right.Leg.Seg.Shank.KneeJoint

So if you can adjust the orientation( ARel property) of these objects then the knee joint can be rotated.

Best regards,
Moonki

3

Hi Moonki,

Thanks for your reply! I have another question if you don`t mind.
It seems that modifying these two objects changes the orientation and position of the knee joint by changing the shank portion only, while the thigh bone stays in place.
I want to change the knee angle by changing the position of the distal femur. How can I do this?

4

Hi,

I’m not sure but also you can change the ‘Main.HumanModel.BodyModel.Right.Leg.Seg.Thigh.Knee Joint’ object.

Then you may be able to change the location and the orientation of this ‘KneeJoint’ node in the Thigh segment.

Best regards,
Moonki

5

Dear all

I have a question concerning this thread:

If I change the orientation of the segments Thigh and Shank in the GaitFullbody model for example, is it then possible to use the same c3d file?

In other words could I simply change the geometries of the bones and use the same c3d data? Because I didn’t perfom any motion caption, for example!!!

thanks in advance for your answers

Arnaud

6

Hi Arnaud,

This is in principle possible, but the results may not be the best.

If you enforce a different bone geometry and have this fixed in the optimization and then allow only the marker locations (some of them) and the motion to be be changed then the model may do this reasonable well provided that the joint locations axes etc are not too different… If there are large difference the model may walk with constantly bent knee if he is too tall compared to the mocap or walking with too streched knees and ankles if he is too short… This will obviously alter the moment arms of the muscles and thus the results…
I know this is not a clear answer, if this approach can be used also depends on what you want to calculate.

Best regards
Søren

7

Hi Søren,

thanks for your answer!

I want to predict the muscle forces necessary to perform the motion defined by the mocap (c3d files) of the GaitFullBody model. So the best way or better approach is to adapt (scale) the bone geometries, that I have, accordingly to the marker locations provided by the c3d file…?

And for Modelling of knocks knees or bowlegs just changing the orientations of the segments (Thigh and Shank) should provide reasonable results…?

8

Hi Arnaud,

If you scale the bones so that is has the same joint centers (and axes) as the scaled model in the GaitFullBodyModel then it should be ok, but then it is no longer the bones you came with so i am loosing the point of doing that?

Concerning the bow or knock legs it would be done changing the knee joint axes, but it would be even more invalid.

I am afraid there is no quick and nice fix to this type of problem, if you follow the approach you have outlined the model may give you some results but these will probably only be qualitative…

Best regards
Søren

9

dear Søren,

yes… of course you are right…
And why should I scale my bones accordingly to the marker locations provided by the c3d file, instead of simply use the scaled model in the GaitFullBodyModel? It will be like reinventing the wheel!!!

thanks for your answers!

cheers!

Arnaud