coordinate system of the sensors

[Hans-Jochen Edenhart]

hello all,
i have my data from my IMUs converted so far. i am trying to find the correect coordinate system.
but the movement goes always in the wrong direction, no matter how i transfer my coordinate system. i really have tried every possibility.
i would be very thankful, if you could tell me, which direction the singel sensors have in your model.
i hope there is someone who would like to support me.
best regards
hajo

[jimmy d]

Hi, Hans-

We had found some bugs in the initial correction heading that may have contributed to your issues. Are you still working on getting your IMUs tracked in OpenSim?

Thanks,
-James

[Henrike Hessling]

Hello guys,

I have the same problems in finding the correct axis direction. It seems the calibrated model acts not as thought when I transform my coordinates in the right way (have it proved). The base heading axis and label are not set by me, because they act strange even more.

Further, the tibia_r body knode BF_ at gastroc_r shows only a half body (half cylinder) with a single y-part. Maybe this is a bug.

Did you made similar experiences?

Thanks,
H.H.

[jimmy d]

Hi, Henrike-

We have been working on fixes to the bugs and they will be included in the soon-to-be-released OpenSim 4.1 beta. We can include you in the beta testing group if you would like to test the fixes?

Cheers,
-J

[Henrike Hessling]

Hello James,

that would be really nice , what do I have to do for joining the beta testing group? and where do I find it?

In case of the BF_at_gastroc_r, I set it from quadrant "y" to "all", but I could not recognize any changes in the model.

Thanks,
H.H.

[jimmy d]

Further, the tibia_r body knode BF_ at gastroc_r shows only a half body (half cylinder) with a single y-part. Maybe this is a bug.

This is not a body, it is a WrapObject. The behavior of the wrap object seems to be what it is designed for so I am not sure what the problem is?

[Henrike Hessling]

Hello James,

thank you for your nice support, you were right. I found out, that the BF_at_gastroc_r was not the main failure. It's the MotionType of the Custom Joints which have to become updated for using in OpenSim 4.0. like it's written here: https://simtk-confluence.stanford.edu/d ... penSim+4.0

Nevertheless I'm interested into the beta version. Is there a chance to get the source code of opensense anywhere?

Thx,
H.H.

[emmanuel ayad]

Hello folks !
Great tool you developed here!
I am trying to make OpenSense working for our IMU system but we are still confused with the good placement of the IMU sensors.

For your information, we are working with the Motesque brand IMU (acc, gyr and mag data). We developed our own algorythm to do the sensor fusion and get the quaternion data.


Sorry if my question seems stupid but we have been stuck for a couple of days now...
My question is, let's assume you set the torso imu sensor with the following axis direction :
+z = forward
+y = right side
+x = vertical

I suppose you have to follow the same positionning for pelvis, thigh and tibia.

But how about the feet ? We are very confused about the right positionning, the feet never go in the forward direction in the simulation. More, do you position the sensor on the heel or like we did, over the shoelaces ?

Looking forward to hearing from you !
Best regards,

Emmanuel AYAD