Velocity data integration giving noisy results

Hello,

I'm trying to integrate velocity (using cumtrapz) data to get the displacement, but when I do so i get weird noise in the signal. I've filtered the velocity data and removed drift with the zero-phase filtering (filtfilt function), but it just doesn't seem to be right. Any ideas?

Cheers,

Cameron

[dpb]

You'd have to attach the .mat file with the original data for anybody to be able to do anything -- and explain what is in it in a 3D array...

You can start, of course, by plotting the time histories to see what they look like...

[Cameron Sprent]

Have edited the question to include the data files. You can see the time history plots in the attached Jpgs.

[dpb]

Something happens in your velocity between beginning and end of the trace...I didn't try to figure out where, but results I see look reasonable for the data as it is...there may yet be a very slight offset at the end although I didn't really investigate that much, either, after I saw the last frame of velocity data, didn't seem worth pursuing further until that is resolved -- if it is an issue -- if the input forcing is, in fact, step-valued, then I suppose it could be what it should be. With no idea what we're looking at/for, it's hard to say anything.

>> displ=cumtrapz(cropped_time,detrend(cropped_vel));
>> figure
>> plot(cropped_time,cropped_vel)
>> hold on
>> plot(cropped_time,displ)
>> xlim([20 20.1])
>> xlim([cropped_time(end)-0.1 cropped_time(end)])
>> close
>> figure
>> subplot(2,1,1)
>> plot(cropped_time,cropped_vel)
>> hold on
>> plot(cropped_time,displ)
>> xlim([20 20.1])
>> subplot(2,1,2)
>> plot(cropped_time,cropped_vel)
>> hold on
>> plot(cropped_time,displ)
>> xlim([cropped_time(end)-0.1 cropped_time(end)])

[dpb]

Added another explantory graph --

Same results as the above for top figure except blown up displacement on RH axis. NB: the non-zero mean; detrending was applied over the full time trace; the mean for the first frame isn't zero as it doesn't match overal mean.

The second is applying detrend to the frame itself; then

>> mean(displ)
ans =
  -7.5495e-06
>>