How to merge two arries with different elements?

I am working on two series of observational data. Both of them are acquired at the same location at the same time only with two different sensors. Now due to the unstable observational environment, both of them would have some temporary failure in acquiring data. Now I would like to merge the observation data of these two sensors by time, so I could fill in the data blank over the whole observation period.

For example, the functional time of sensor one is [0,1,2,3,4,6,7,8,11,15,23] (hour of the day), while the sensor two are functional at [2,3,4,5,7,9,10,12,13,14,16,18,20,22]; How should I merge these two time record as one, with several commands?

the expectation on the merged array is [0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,18,20,22,23];

I realize the algorithm is simple, but I would like to have some related function in matlab that I can apply on this issue.

Thanks in advance!

Jason

[Adam]

Which sensor is expected to take precedence when both have readings or do you want to average them?

[Shenjie Zhou]

Hi Adam,

Thank you for your comment. I am glad my question makes sense to you. I would like take the average when both sensors have the readings.

[Adam]

Also what happens if neither sensor picked up a reading at a given time?

[Shenjie Zhou]

Fill up the time record, but put the corresponding readings as NaN.

Answer 1

[Adam]

t1 = [0,1,2,3,4,6,7,8,11,15,23];
r1 = rand( 1, 11 );
t2 = [2,3,4,5,7,9,10,12,13,14,16,18,20,22];
r2 = rand( 1, 14 ) * 100;

rNew = NaN( 2, 24 );

rNew( 1, t1 + 1 ) = r1;
rNew( 2, t2 + 1 ) = r2;
rNew = nanmean( rNew );

should do the job, obviously with your real data values instead of the random data I used.

You can also use

rNew = mean( rNew, 'omitnan' );

if you prefer.

Obviously the time array you can just create as:

tNew = 0:23;

[Shenjie Zhou]

That is brilliant, thanks a lot!

Answer 2

[Steven Lord]

If you're using the new timetable datatype introduced in release R2016b to store your data, take a look at the synchronize function. If not, use interp1 to interpolate both signals to a common time basis.

[Shenjie Zhou]

Hi Steven,

Interpolation is also a solution, but one of the sensor has missed records more than the other one, so the interpolation on this sensor readings may introduce some error. I am trying to keep the observational data as original as possible.

Thank you for you comment!