Meta Motion R - Displacement tracking through Sensor Fusion

NHK

Hi guys.

I don´t know if this question has been answered already, but here goes. 

Do you guys have any best method for calculating displacement using Meta Motion R? In your Facebook advertisement, you post, that the board is optimal for "accurate displacement calculation". I guess that this is not done by double integrating acceleration signal due to signal noise multiplication.

So this is done with the sensor fusion method which obtains quaternions. 

Can anyone help me with best practices for calculating axis displacement through quaternion if possible? Links to any tutorials, documentation etc?

I am trying to calculate changes in Z-axis down to the centimeter if possible. So Barometer(altitude) is out of the question. 

So far if have only tried double integrating a highpass filtered dataset but without any luck.

Regards

Nils Kristensen

Guarneri

Hi Nils,

We have not tested any displacement calculations over a long period of time, but perhaps something like this would help:

https://math.stackexchange.com/questions/90081/quaternion-distance

Thanks,

Yu

kevihead

I've been using Madgwick's gait tracking algo for a similar thing. I don't use sensor fusion myself, but I tried it once and it worked reasonably well (but not as good as using his algo with raw accelerations and rotations). You'll have to take out half of the stuff in the main script, but it should work. 

Depending on the length of the movement and the nature of it, you might not get it to within a cm as sensor fusion's accelerations are offset and it's difficult to correct them.