Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Vector integration - accelerometer data processing

295 views
Skip to first unread message

Sebastien Moindrot

unread,
Mar 30, 2001, 10:31:10 AM3/30/01
to
Hi there,

I'm looking for a function to help me to process the output from an
accelerometer to obtain velocity and Position. The data are
acceleration measured every 0.1s. If such a function is available
could please tell me which one it is or if you have met such a problem
in the past.
Thanks for your help.

Sebastien

Amphlett

unread,
Mar 30, 2001, 2:19:33 PM3/30/01
to
You need to high-pass filter your acceleration data (suggest: 10% of nyquist
butterworth high-pass) and then use cumsum()*dt and cumsum(cumsum())*dt*dt.

e.g.

% I may have the arguments round the wrong way here.
[b,a]=butter(10,0.1,'high');

acc=filter(b,a,yourdata);
vel=cumsum(acc)*dt;
pos=cumsum(vel)*dt;

John Lukesh

unread,
Mar 30, 2001, 11:50:38 PM3/30/01
to
Hi,

I think you should look at MATLAB’s ode (ordinary differential equations)
functions, function ode45 for example. You should be able to set up your
integration of accelerometer data problem as a set of first order
differential equations and use one of the ode functions to solve for
velocity and position. Go to help/function/ode45 as a start.

Regards, John Lukesh

Jun

unread,
Dec 11, 2012, 7:28:14 PM12/11/12
to
John Lukesh <jlu...@deltanet.com> wrote in message <3AC566B5...@deltanet.com>...
> Hi,
>
> I think you should look at MATLAB?s ode (ordinary differential equations)
Dear John,

Can you tell me your new email address please?

Thank you very much!

Cheers,

Jun

Steven_Lord

unread,
Dec 12, 2012, 9:32:32 AM12/12/12
to


"Jun " <wang...@163.com> wrote in message
news:ka8j2u$q05$1...@newscl01ah.mathworks.com...
Considering that John's last posting was over eight years ago (according to
MATLAB Central) if you want to ask a question you should probably post it to
the newsgroup rather than trying to email him. [Please do NOT email it to me
directly either; post to the newsgroup so that everyone can read it and
perhaps offer suggestions.]

That being said, ODE45 or CUMTRAPZ would be good first steps in integrating
accelerometer data; use ODE45 if you have a function you've fit to the data
that fits well/matches the theory behind your experiment, CUMTRAPZ if you
have just the data itself.

--
Steve Lord
sl...@mathworks.com
To contact Technical Support use the Contact Us link on
http://www.mathworks.com

TideMan

unread,
Dec 12, 2012, 1:54:25 PM12/12/12
to
Furthermore, this topic has been discussed many times before in this forum and various methods have been suggested. Use the search facility to find them.
0 new messages