simtk.org

Hi all,

I'm currently performing RRA and I'm getting an odd result where the force residuals are low (<5N) but the moment residuals are still high (>80Nm). I've gone through the process of iterating the model and updating the mass of the segments. Changing the COM locations in the actuator files actually seem to make the residuals even higher (when compared to using the default files).

Any Ideas?

Thanks,
Alex

RRA changes the kinematics, which can affect the relative location of the foot and the GRF — changing the residual moment.

Hi, James.

So do you have any idea or advice how to decrease these bad residuals (force>25N,moment>75Nm )?

Hi, Bradley.

Have you sorted out your problem? I encountered the similar problem (residual vertical force much greater than the suggested one but the moments are OK). Looking forward to your reply.

Thanks in advance,
Linjie

I noticed both the tutorial models gait2354 and gait2392 have ControlConstraints files like gait2354_RRA_ControlConstraints but i cannot find where to input it. Is it helpful to improve the residual results?

In case you haven't seen them already, there are some troubleshooting tips at the bottom of the "Getting Started with RRA" page in the Confluence documentation (http://simtk-confluence.stanford.edu:80 ... d+with+RRA).

Thanks, Tom.

And yes, I have seen them before and tried most of those tips except one setting to make sure "use_fast_optimization_target" is false. I dont know how to find that setting file.

thanks,
Linjie

I still could not find the setting of "use_fast_optimization_target" but successfully add the constraint file to my RRA analysis through manually copy and paste the file name in the RRA setting xml file.

However, so far I still haven't got my residual force result (vertical direction) right, even though I have tried the adjustment of excitement control values and optimal force values. I noticed the drifting of pelvis was indeed improved by doing modification of the control and optimal force values, but the pelvis jumping is still existent, which I could observe the action location of the ground reaction force is actually away from the correct one (the action locations are right through associating GRF data with the motion data). I am quite confused that I did set the force on the calcaneous bone, and the force express and I also tried the modification of pelvis mass and its location, nothing changed better.

Is there any suggestion how I could fix the wrong action location of ground reaction force or what any other potential problems I should check.

Thanks,
Linjie

attached pictures are the GRF acting on the wrong location, drifting away actually with the jumping movement of pelvis during the motion of lower extremity.

The problem of pelvis drifting has been sorted by changing the direction of gravity acceleration. And my oversized residual force in the vertical direction was reduced to suggested range by increasing the mass of pelvis. I dont know if this is allowed, for I dont have the torso added in my model.

It seems like you are only modeling the pelvis and one (right?) leg; what are you doing to account for the mass of all the other bodies (trunk, arms, head, other leg). If you are analysing walking gait, what are you doing to take into account external forces when the other leg is in contact with the ground.