Accurate position detection with the new 101

Does anyone know: with the new Arduino 101 running the Madgwick filter code, is it possible to derive board position (not orientation)? IOW, if you move the board 1 inch to the right, can this be detected? If you move it 1 foot forward, can this be detected? Etc.
Thanks!

Since it contains an accelerometer, that should certainly be possible, although I know nothing about the Madgewick filter stuff.

Bear in mind that accelerometers, when computing distance and direction, are integrative. That means that each new estimate builds on the last, so any error that has accumulated will be carried forward (although it may also accidentally partially cancel out).

What this implies is that while you'll probably get very accurate results for moving an inch, or a foot, the accuracy will decrease as the distance increase. So don't count on a reliable reading at say, 1/2 mile.

kevykk:
running the Madgwick filter code, is it possible to derive board position (not orientation)?

The Madgwick paper is named “An efficient orientation filter for inertial and inertial/magnetic sensor arrays” so I’m guessing ‘no’.

The basic problem with this position measurement is you're on board a dragster car in the middle of a 7-second pass.* If you aren't aligned perfectly, then errors go up to KILOMETERS in just a minute. There is unfortunately no such thing as "just measuring an inch" with accelerometers.

  • = Otherwise known as "gravity."