??? Subscript indices must either be real positive integers or logicals. Using findpeaks

I've have been using findpeaks in MatLab to locate the maximum and minimum points of a waveform with no problem, but in the last 20 minutes or so the error:
??? Subscript indices must either be real positive integers or logicals.
Has appeared an I have no idea why. Even trying simple exercises with test data has resulted in the same error. For example if I were to have the dataset:
test = [ 0.1 0.5 0.9 0.5 0.2 0.6 1.0 0.7 0.3 0.1 ]
and used the code:
peaks = test(findpeaks(test));
I would expect the result:
peaks = [0.1 0.9 0.2 1.0 0.1 ]
but for some reason this is no longer the case.
Please advise.
Thanks,
Jared.

2 Comments

Where does it say that the problem is occurring? Which line, which file?
The problem is occuring at: peaks = test(findpeaks(test));
This code previously returned maximum and minimum peak values in a single matrix does now is problematic

Sign in to comment.

 Accepted Answer

findpeaks returns the peak, which is 0.9, if you don't accept both output arguments of it. 0.9 in not an integer and so it cannot be used as as index into test. Since you want the peaks, and that's what findpeaks returns, you can simply do this:
testData = [ 0.1 0.5 0.9 0.5 0.2 0.6 1.0 0.7 0.3 0.1 ]
peaks = findpeaks(testData)
and get the results:
testData =
0.1000 0.5000 0.9000 0.5000 0.2000 0.6000 1.0000 0.7000 0.3000 0.1000
peaks =
0.9000 1.0000
These are the peaks, but this does not match what you're calling peaks. It looks like you're including the valleys in how you define peaks. In that case, you might want to run findpeaks again on the inverted data (subtract your data from the max of the data).
maxPeakValue = max(peaks);
% Surround with 0 so you can get the valley
% should it occur at the first or last element.
invertedData = [0, maxPeakValue - testData, 0]
valleys = maxPeakValue - findpeaks(invertedData)

More Answers (2)

You've probably declared a variable called findpeaks by accident. Try
which findpeaks
and confirm it's actually pointing to a function

3 Comments

clear findpeaks
to remove the variable overwriting it
which findpeaks returns:
C:\Program Files (x86)\MATLAB\R2011a Student\toolbox\signal\signal\findpeaks.m
and I don't have anyother variables in my current workspace other than test and peaks
I've tried clear findpeaks and running the code again but the problem remains :(

Sign in to comment.

[pks,locs] = findpeaks(test);
pks
test(locs)
pks contains the peak values
locs are the indices
test(locs)

5 Comments

Is there a way to implement this to return the max and min points?
you mean the maximum and minimum peak?
sort(findpeaks(test))
If you just want the min and max elements of the vector, then just use min(), max()
Sorry I was not clear, the dataset I will eventually be using represents a repetitive exercise over a 30 second period. I need every high peak and every low peak as they represent the beginning and end of an exercise.
He wants the valleys. In my answer I recommended Jared invert the data and run findpeaks again. Valleys will turn into peaks once you've inverted it.
haha, I blame TMW for this confusion (pks vs locs). findpeaks suggests that it returns indices a. la. find, you'd actually want it to be called peaks. But we can't have that, can we!

Sign in to comment.

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!