AddBiomechanics vs. Traditional Scaling

Upload mocap files and get an optimally scaled OpenSim model and IK back in minutes. Share your data with the community. Browse and download biomechanics data. Hosted by Stanford University.

Available at https://addbiomechanics.org
POST REPLY
User avatar
Kaitlyn Downer
Posts: 11
Joined: Tue Jul 11, 2023 8:47 am

AddBiomechanics vs. Traditional Scaling

Post by Kaitlyn Downer » Mon Mar 04, 2024 5:57 am

Hi,

I am trying to scale one subject (Female, 1.645 m, 89.5 kg) using AddBiomechanics and got different results than when I used the traditional scaling method in OpenSim. I am using the gait2393 model with a custom marker set (12 markers: RASI, LASI, SACR, RKNE, RANK, RHEE, RTOE, LKNE, LANK, LHEE, LTOE, C7)
AddBiomech.png
AddBiomech.png (462.49 KiB) Viewed 369 times
AddBiomech2.png
AddBiomech2.png (430.95 KiB) Viewed 369 times
1. When looking at the logs, the found skeleton mass was found to be 95.7383 kg. Is this much of a difference common?

2. Why is the pelvis so small in the AddBiomechanics model? I understand that the other scaling resulted in a larger than normal pelvis, but the mass of the AddBiomechanics model's pelvis is actually 1 kg more than the other.

3. Is the force status displayed on the results the max or RMS? Essentially, what can I look at to compare to the thresholds previously described in RRA to determine the RRA results.

4. I am concerned because the height is not the same in the two. What could cause this?

5. Does AddBiomechanics require more markers than I am using? This is a reduced marker set that we used for scaling, but the experimental data does contain more markers.

Thanks,
Kaitlyn
Top
User avatar
Nicholas Bianco
Posts: 1027
Joined: Thu Oct 04, 2012 8:09 pm

Re: AddBiomechanics vs. Traditional Scaling

Post by Nicholas Bianco » Fri Mar 08, 2024 11:16 am

Hi Kaitlyn,

I think most of the issues here stem from not having enough markers. 12 markers just won't be enough to describe the whole body's geometry. To answer your specific questions:'

1. Differences are expected, but this does seem to be a bit larger than normal. I would double check how you are pre-processing the ground reaction force data since the dynamic fitting process adjusts masses based on the provided forces.

2. The three markers on the pelvis are probably not giving you enough y-direction information, which is causing it to be "squished" like it is.

3. Those values are the magnitude of the residual forces and torques averaged across the motion.

4. This is probably due to lack of marker data in the upper body.

5. Answered above.

Best,
Nick
Top
POST REPLY