# Subtracting centrifugal acceleration from acceleration caused by movement

I've been playing around with dead reckoning stuff on an rc plane by attaching a gps, 3 axis gyroscope, and 3 axis accelerometer to the plane.

When the plane isn't turning, my algorithm works pretty well however when I enter a turn, the readings get way off. Specifically, the plane appears to be moving at a faster velocity than it actually is thus all turns are overshot.

I believe I see this due to centrifugal acceleration and would like to cancel it out, but I'm getting quite lost, especially with 3d vectors. How might I go about subtracting this observed acceleration from acceleration caused by movement?

Thank you

Related Other Physics Topics News on Phys.org
Let me add what I have identified so far.. for at least identifying what I believe to be the acceleration component I need to subtract:

F = ma$_{c}$ = (mv$^{2}$)/ r = mrw$^{2}$

The definitions are standard as of this wiki page:
http://en.wikipedia.org/wiki/Centripetal_force

I know:
(gps) speed = v
(gyroscope) angular velocity = w

I can calculate r via:
(mv$^{2}$)/ r = mrw$^{2}$
v$^{2}$/ r = rw$^{2}$
v$^{2}$ = r$^{2}$w$^{2}$
r$^{2}$ = v$^{2}$ / w$^{2}$
r = $\sqrt{v^{2} / w^{2}}$

With r and w known, I can now calculate $a_{c}$ as follows:
m$a_{c}$ = m$rw^{2}$
$a_{c}$ = $rw^{2}$

Am I anywhere close to being on the right track? :rofl: