Questions about static optimization

Provide easy-to-use, extensible software for modeling, simulating, controlling, and analyzing the neuromusculoskeletal system.
POST REPLY
User avatar
Yanran Jiang
Posts: 10
Joined: Sun Feb 09, 2020 11:48 pm

Questions about static optimization

Post by Yanran Jiang » Fri May 29, 2020 8:09 am

Hi there,

I'm doing simulation of squat. When I run the static optimization, there are some messages I don't understand well.

1. " Try appending the model with additional forces or locking joints to reduce the following acceleration constraint violations pelvis_tilt: constraint violation=2.919 "

I have checked my Inverse Dynamics result, the motion and joint force look reasonable and there is no sudden jump values. So is the only solution to add additional actuators/force set? If so, how should I know what kind of actuator should I add? How to design it? Could I apply the actuator file of walk(e.g gait2354_CMC_Actuators.xml)?

2. "Optimizer failed : Ipopt: Maximum iterations exceeded"

Is it right just to increase maximum iteration values in the set up file?

3. "unable to achieve required assembly error tolerance"

I have no idea what is it?

4. "Performance = 480"
What's the meaning of the performance at each time instance?

Thank you in advance
Attachments
sssss.PNG
sssss.PNG (17.61 KiB) Viewed 751 times
aaa.PNG
aaa.PNG (17.76 KiB) Viewed 751 times

Tags:

User avatar
Michael Asmussen
Posts: 67
Joined: Mon Jul 11, 2016 7:46 am

Re: Questions about static optimization

Post by Michael Asmussen » Fri May 29, 2020 11:43 am

Hi Yanran,

I can provide some input on these questions.
1. " Try appending the model with additional forces or locking joints to reduce the following acceleration constraint violations pelvis_tilt: constraint violation=2.919 "

I have checked my Inverse Dynamics result, the motion and joint force look reasonable and there is no sudden jump values. So is the only solution to add additional actuators/force set? If so, how should I know what kind of actuator should I add? How to design it? Could I apply the actuator file of walk(e.g gait2354_CMC_Actuators.xml)?
It appears that there is a violation with the pelvis tilt degree of freedom. You could put a torque actuator on that degree of freedom and see if that solves the issue. You could start with a really high torque value and then reduce it down as low as you can until the solution fail. Ideally this torque actuator value will be quite low. Also, check that data around 17.22 to see if there is an anomaly in your data.
2. "Optimizer failed : Ipopt: Maximum iterations exceeded"

Is it right just to increase maximum iteration values in the set up file?
You could try to increase the iterations, but it still might not solve the issue. My suggestion would be to start with the torque actuator on the pelvis tilt degree of freedom.
3. "unable to achieve required assembly error tolerance"

I have no idea what is it?
My understanding is that there is an error tolerance between the moment generated from inverse dynamics and the moment produced from the optimizer using the muscle moment arms and estimated forces. The solution you are producing at the end of your iterations is larger than this error tolerance.

I hope this helps!

Mike

POST REPLY