Accelerometer Accuracy

Iain B. Findleton ifindleton at
Sat May 2 15:14:14 CEST 2009

After many experiments with the accelerometers, I  notice that there is,
on my FR, a considerable difference between the readings from the 2
devices. Aside from the noise issue, the difference between the measured
gravity between the 2 devices is of the order of 1m/s**2, device 0 being
lower than device 1.

Is there any way of calibrating the output from these devices? Simple
noise suppression using thresholds appears to aggravate the problem. In
my tests, the error computing the angles between the local gravity
vector and the device axes when the FR is at rest on a flat surface
appears to be in the range -1.5 to +1.5 degrees, something that would
appear to make these devices more or less useless for anything beyond
very crude estimates of local acceleration or gross changes of position.

Iain B. Findleton
Tel: 514-457-0744

More information about the community mailing list