Thread Subject: double integration of accelerometer data

Hi,

I'd like your assistance in integration an accelerometer data to get position. i've read other responses but they do not seem to help.

I know there are problems of accelerometer drift and initial conditions attached to the acceleromet output. I applied the FFT to the signal before integrating(using cumtrapz) but it still does not help out with the drift. the signal output remains the same.




 

On Feb 2, 3:53 am, "kunle " <geedee...@yahoo.com> wrote:
> Hi,
>
> I'd like your assistance in integration an accelerometer data to get position. i've read other responses but they do not seem to help.
>
> I know there are problems of accelerometer drift and initial conditions attached to the acceleromet output. I applied the FFT to the signal before integrating(using cumtrapz) but it still does not help out with the drift. the signal output remains the same.

Your post is so vague, I doubt if anyone can help.

When you say "I applied the FFT to the signal before integrating" what
does that mean?
Do you mean that you integrated the resulting Fourier transform with
cumtrapz?
If so, then you need to re-think this.

To remove the drift, you need to high-pass filter either before or
after integrating.