Issue Recognising Umberger Metabolic Probes in 4.0

[Aaron Fox]

Hi All,

I'm having some issues with CMC not recognising the metabolic probeset I have in my model, in that when CMC runs it starts up giving out a warning for each muscle (e.g. WARNING: Umberger2010MuscleMetabolicsProbe_MetabolicMuscleParameter: Muscle 'psoas_r' not found in model. Ignoring...), and subsequently the probe reporter doesn't give out any results. I am using some existing code that did work in 3.3 and am wondering whether there are any differences with how these are added/written into the model?

I've been able to create the probe set using relevant commands in Matlab (i.e. Umberger2010MuscleMetabolicsProbe, set_probe_operation, set_report_total_metabolics_only, setName, setEnabled and addProbe), and I believe this is recongised - hence I think the error is in how I've added the specific muscles components into the probe set (I've done this using the addMuscle command). I've noticed if I open the model in the GUI after doing this those that the probe set seems to have no components (listed as 'empty').

The model file is apparently too large to attach here so I've attached a screenshot of what the probe set code looks like to hopefully get some clarity on whether there is an issue there.

Thanks,

Aaron

[jimmy d]

The section of code that does the check seems to be here

It does sound like the the 'live' model is not getting updated to have the probes in it (hence the list being 'empty'). Does the code perform an .initSystem() after it has added the probes? perhaps if you post a folder with the example code and a model we could debug.

-j

[Aaron Fox]

Thanks for the point in the right direction James. It seems the response of that section of code is to turn the probes off, which looks like what is happening. I don't think I have an initSystem() prompt after adding the probes so I will add and see if it resolves. If not I'll post some extra code for some debugging help.

Aaron

[Dmytro Radevych]

Hello,

I have the same issue with a model I'm using. As I run through CMC I get the same warning for every muscle in my model:

WARNING: Umberger2010MuscleMetabolicsProbe_MetabolicMuscleParameter: Muscle 'addbrev_r' not found in model. Ignoring...

I have added the metabolic probes with python script from tutorial Gait10dof18musc and I am using the Rajagopal2015 model.

Can I rely on the metabolic calculations, which I get from CMC, in spite of all warnings? Or do I need to debbug the model somehow?

Dmytro

[Aaron Fox]

Hi James,

As an update to this I'm still having the same troubles with the connection of the probe muscles to the model muscles. I've tried using the .initSystem() command as well as the .finalizeConnections() command but to no avail. The issue seems to be, as you've pointed out, that at that point in the code the name of the probe muscle (e.g. addbrev_r) doesn't match up with any muscle names in the model. Weirdly though, if I load the model into Matlab and use the same command (i.e. OsimModel.getMuscles().getIndex('addbrev_r')), it outputs an index value that should align with matching these up in that code (i.e. it is 0 or greater).

Despite it seemingly all matching up I'm wondering with the update to 4.0 whether there is some difference with the naming conventions for muscles here that I need to change when allocating the metabolic probe parameter set? All the tests I can think of running suggest that there are muscles called 'X' in the model, and this is what they are called in the probe parameter set, but they don't seem to be talking to one another when running the analyze tool.

Any suggestions would be much appreciated!

Thanks,

Aaron

[Scott Brennan]

I'm having the same issue running the AnalyzeTool with the Metabolics Probe attached. I'm using the LaiArnold2017 model where I have included the muscles from my model in the same probe format as the Gait10dof18Musc example. However, when I try to run one of the tools I get this warning for every muscle: "WARNING: Umberger2010MuscleMetabolicsProbe_MetabolicMuscleParameter: Muscle 'vaslat140_l' not found in model. Ignoring...". The names match up but I haven't figured out a way for it to recognise that the muscles in ForceSet are the same as ProbeSet.