Hi Matteo, I agree with your explanation in that the direction of the gravity vector as measured by the accelerometer is infact upwards. It is important to be aware that under quasi-static conditions the assumed direction does not effect the calculation of inclination as long as the user is consistent in their calculations. In the next release of the 9DOF Calibration application we will reverse the direction of the reference vector. Thankyou for your assistance on this topic.
Best Regards, Karol From: [email protected] [mailto:[email protected]] On Behalf Of Matteo Giuberti Sent: 07 September 2012 16:47 To: Karol O'Donovan; [email protected] Subject: Re: [Shimmer-users] SHIMMER calibration Dear Karol, thank you for the quick reply. Actually, I think the idea behind an accelerometer is much more involved. Just to go step-by-step, I want you to reason upon this: when an accelerometer is in free-fall (imagine you let it fall from your hand to the ground) what will it measure? Maybe I'm wrong, but I always thinked that it will measure 0g. If this is right, then it is not actually measuring the gravity force but the accelerations that act on the device apart from the gravity force. Indeed, my previous idea (that of the + and - signs in the previous mail) was based on the fact that if a device was leaned still on a table it should measure an acceleration that counteracts that of the gravity force (even if physicians will tell me that it is actually a force and not an acceleration) in order to let the device still! So, I think that a SHIMMER whose z-axis is pointing vertically up should measure a +1g (or +9.8m/sec^2) AND NOT -1g (that is obtained instead through your calibration). If there is something wrong in my reasoning please try to let me understand! P.S. Just to provide you more evidence I just found on wikipedia this: http://en.wikipedia.org/wiki/Accelerometer If you read in the introduction and in the "physical principles" sections you will find more or less the explanation I just provided to you. I know Wikipedia is not exactly a trustworthy source, but it actually said what I knew about that topic! Regards, Matteo 2012/9/7 Karol O'Donovan <[email protected]> Dear Matteo, The gravity vector is a vector of magnitude 9.8 m/s^2 which points downwards (i.e. towards the centre of the earth). Using the convention suggested in the manual, when you have the Shimmer orientated such that the sticker/leds are pointing upwards then +z-axis is also pointing upwards. The gravity vector is thus pointing in the opposite direction to the +z-axis and therefore registers a negative value. Hopefully this helps to resolve the problems you are having. Best Regards, Karol From: [email protected] [mailto:[email protected]] On Behalf Of Matteo Giuberti Sent: 07 September 2012 11:35 To: [email protected] Subject: [Shimmer-users] SHIMMER calibration Dear all, I have some SHIMMERs with the 9DoF module and I have to calibrate the sensors (accelerometer, gyroscope, and magnetometer) in order to use their output in an orientation estimation filter. I'm using the calibration software provided to the member user and, reading the manual, I've noticed that the instructions provided to calibrate the accelerometer lead to a calibration of the accelerometer that is different from all the previous devices that I've experienced in the past. Let me first say that I've used the same coordinate axis convention that you suggested in the manual. Following that convention, I was expected to see, for example when the Shimmer main side (i.e., the one with the sticker and the leds) is pointing toward the ceiling, a z-acceleration almost equal to +9.81 m/sec^2. Instead it is -9.81 m/sec^2. Can you motivate me this choice (that, I repeat, is different from the conventions that I previously experienced in other devices)? Do you think it will be a problem to calibrate ONLY accelerometer using my convention? I hope I clarified my point. Otherwise, I'll try to explain it better. Best regards, Matteo
_______________________________________________ Shimmer-users mailing list [email protected] https://lists.eecs.harvard.edu/mailman/listinfo/shimmer-users
