How do I interpolate a set of data and change the timestamp to specific points in time?

I have several sets of data coming from different sensors which register data at different time-intervals. I want to synchronize the time of all the data down to milliseconds, and I want to interpolate the values for the timestamps so they match the new timestamp.

Some sensors register a new value every hour, while others register for every 10th, 8th, 30th, etc, minutes.

[Walter Roberson]

timetable() objects together with retime() might be suitable for you.

[Anders Teigmoen]

I have come to the same conclusion. One problem though. I keep getting the error: ''Error using timetable/retime (line 103) Input timetables must contain sorted (increasing or decreasing) row times when synchronizing using 'spline'.''. I have already sorted the timetable by applying ''sortrows(TT1)'', but I still get the error above. Do you have any idea why this error occurs?

[Peter Perkins]

Anders, I think we'd have to know what;s in your times. Are there duplicates?

You may need to use retime to get rid of those before interpolating -- retime to a target time vector that's unique(tt.Time), and reconcile the multiple values using 'mean', 'max', or whatever.

[Anders Teigmoen]

I managed to solve the issue. There were no duplicates, but a NaN lurking in my data. Thank you for the response.

Answer 1

[KSSV]

Read about interp1. With this you can get all the data into same time stamps.

An example:

t1 = (0:10:100) ;
y1 = rand(size(t1)) ;
t2 = 0:5:100 ;
y2 = rand(size(t2)) ;
plot(t1,y1,'.r') 
hold on
plot(t2,y2,'.b')
% interpolation 
ti = 0:1:100 ;
y1i = interp1(t1,y1,ti) ;
y2i = interp1(t2,y2,ti) ;
plot(ti,y1i,'r')
plot(ti,y2i,'b') ;