simtk.org

Hi Everyone,

I am running static optimization on a jump landing task and my resulting muscle forces are extremely noisy. I have included a plot of the vastus lateralis with the GRF as a comparison. As you can see, my GRF are clean and I have used RRA on the kinematics reducing the noise in them as well. Has anyone experienced this noise before? If so, were you able to identify the cause?

Are you using a muscle analysis or are these the direct outputs of SO?

Are all muscle forces just as noisy? Could you post your SO setup file?

Hi Christopher,

Your GRF data looks clean. Since you are using outputs from RRA, I'm assuming that your kinematics data have already been filtered.

that said...

How do the moments about the knee look? If it were an issue with the CoP or force assignments, presumably the the moments would look very wrong as well... You could check the CoP data by plotting that position relative to markers on the foot.

Thanks everyone for your suggestions.

James - These are direct outputs of SO. Although I have run a muscle analysis, I am still rather unsure on how to effectively use the outputs. As this is one of my goals this week is to develop a better understanding of muscle analysis, do you have any literature I can read as a starting point? I.e. what are important things I should be considering when examining the outputs?

Christopher - Yes, all muscles are just as noisy and I have attached my setup file. I've also included my SO actuators file that contains my reserve actuators parameters. The noise tends to decrease as I make the reserve actuators cheaper but I want to refrain from making the reserves extremely cheap for obvious reasons.

Derek - Yes, the kinematics have been filtered. I have also attached a plot of my knee moments, which are also quite clean. I have not investigated the COP data but will try to extract these values from the GRF files.

Kenneth:

Thanks for attaching those files. They look fine to me. I wonder if the ligaments in your model could be causing the issue. Could you try running static optimization on a model that doesn't contain ligaments to see if that causes the noise to go away?

Hi Chris,

Thank you for the information. Yes, the issue was the ligaments as they were too stiff so the SO was failing to find a solution at various time points, which resulted in the large spikes. Muscle forces are smooth when the ligaments are loosened up.

Thanks everyone for your help!

-Kenneth