Correction of measurement data by interpolation

Hi!

I have some corrupt measurement data, which I want to correct. The data acquisition system didn't get an actualized value in the correct time interval. It should get a new value after 0,2s. It happens, that I get the next value e.g. after 10s. My recorded data shows then 10s the same value and afterwards a big step to the next value. I need a script which recognizes where this happens and it shall correct it by interpolation.

Can you help me?

Kind regards

Andreas Wagner

[Star Strider]

Please attach the code you use for importing that file and creating the vectors you are plotting.

It obviously requires xlsread, however it has such a strange format that I have no idea how to use the result.

[Adam Danz]

Sounds like diff() and interp() or interp1() are the primary tools you'll need.

[jonas]

The data you attached exhibit no such behaviour. Consistent steps of 0.2 (see attachment)

@Star Strider: I was able to import it simply by

A=xlsread('data.csv')

[Star Strider]

I could read it easily enough. Making sense of it is impossible.

I want to know how Andreas Wagner read and formatted the data shown in the plot image.

Thus, my Comment.

[jonas]

I'm not sure what has happened to the upload. When I download the file to my computer I get a plain excel-sheet with two columns, from which I can reproduce the figure AW attached. I've attached the data here. When I download the same file to my phone I get something completely different.

EDIT: after a closer look, the content is obviously the same but its formatted in a very strange way on my phone. Probably has to do with excel version?

[Star Strider]

It would be best to upload it as a ‘.txt’ file instead. That avoids version problems.

All I know is that it looked completely weird when I used:

[D,S] = xlsread('data.csv');

I requested the second ‘S’ output because I saw a number of NaN values in ‘D’.

I am using Excel 2016 and MATLAB R2018a.

[jonas]

Yep, same issue on my phone. Took me a while to understand your last comment, as everything worked on my end :)

Anyway, I put the data as a txt in the previous comment. Still, there are no 'skipped data', so not much to work with...

[Image Analyst]

Sometimes you have 3 numbers per line and sometimes 4 numbers per line. Is this one of those numerical formats that use a comma for the decimal point rather than a dot/period? Otherwise, why are there different number of numbers on each line?

[Andreas Wagner]

Comma is the decimal separator

[Andreas Wagner]

I've uploaded the file as .mat.

Kind regards