scaling gait2354 model

[Wayne Johnson]

Hi, I am trying the scaling tutorial using the gait2354.osim model:
https://simtk-confluence.stanford.edu:8 ... e+Dynamics

I followed the instructions for scaling and the model scaled w/out error. I have attached an image of the resulting scaled model (subject01), the unscaled model (3Dgait2354), and the experimental data displayed (blue dots from subject01_static.trc file).

Here are my questions:
1. I assume the offset between the unscaled and scaled model is intentional so it is easy to tell them apart???
2. The scaled model markers appear to have moved only in the yz plane to match the experimental data, but not in the x axis direction. Why is this?
3. I tried running the scaling again using the subject01 model but it still did not resolve question 2. In fact, the RMS marker and total errors increased. Why?

Thanks in advance for your help.

gait2354scaled.PNG (232.52 KiB) Viewed 494 times

[jimmy d]

1. I assume the offset between the unscaled and scaled model is intentional so it is easy to tell them apart???

There is an offset applied to the model after running the ScaleTool(). If you right click on the model in the Navigator and select Display->Model Offset, you will be able to set the offset to zero. Once you have removed the offset, you will be able to observe the overlay between the Model and Experimental markers.

2. The scaled model markers appear to have moved only in the yz plane to match the experimental data, but not in the x axis direction. Why is this?

Im not to sure what you mean by this. Comparing the location of the R.Thigh.Upper marker in the unscaled (0.018 -0.20 0.064) and scaled (0.046 -0.22 0.091) models shows changes in location in all three planes. Unless the Model markers are designated as fixed, they will be moved to overlay the location of the experimental markers.

3. I tried running the scaling again using the subject01 model but it still did not resolve question 2. In fact, the RMS marker and total errors increased. Why?

Because there are Manual Scale factors in the setup file, so each time you perform scaling on the same model you are adding more error to the model.

[Wayne Johnson]

Thanks so much. Once I set the x-offset for the scaled model to zero this addressed question #2.