Find Peaks error - Expected X to be strictly increasing

Attempting to generate a waveform similar to the image, ploting displacement against time

However, getting an error where findpeaks needs x to be increasing

How would I get around this error without altering the displacement data which cant be strictly increasing?

Code Below:

%Averaging I_0

idxI0 = find(I_0_displacement == min(I_0_displacement)); %first response peak index
averageLineI0 =  mean(I_0_displacement(idxI0:end));        %average after 1st true peak
[pks_I0, locs_I0] = findpeaks (I_0_displacement(idxI0:end), I_0_time(idxI0:end),'NPeaks',2, 'MinPeakHeight', averageLineI0); % find the first two peaks x, y values
[nRowsI0, ~] = size(I_0_time);                                   % for adding the average value line

%I = I0:
figure(1)
hold on
plot(I_0_time, I_0_displacement, I_0_time, averageLineI0*ones(nRowsI0,1)); % plot the response and average line
plot(locs_I0, pks_I0, 'ko'); % plot the average line
 %create lines which stem from the average line to the peak values:
[~,I0_c1] = size(averageLineI0:pks_I0(1));                                    
[~,I0_c2] = size(averageLineI0:pks_I0(2));
plot(ones(1,I0_c1)*locs_I0(1), averageLineI0:pks_I0(1), ones(1,I0_c2)*locs_I0(2), averageLineI0:pks_I0(2)); 
title('I0 Displacement VS. Time') % add a relevant title
ylabel('Displacement(mm)') % add y axis label - check that you're plotting positive time, as per the note above!
xlabel('Time(s)') % add x axis label - check that you're plotting displacement in mm
legend('Displacement', 'Average Line', 'peak values', 'x1', 'x2')
axis([0 10 -60 0])   
hold off

Screenshot:

Answer 1

[Bjorn Gustavsson]

Check that your X-variable is actually monotonically increasing (you might have replicate points, which then messes everything up for you):

Increasing = all(diff(I_0_time(idxI0:end))>0)

In the case you have replicate X-values you'll have to resolve that problem. That is simple if the Y-values are identical then just use unique to find the unique X-values and their indices and use those indices to select the corresponding Y-values. If the Y-values aren't identical you'll have to use your best judgement - perhaps taking the average?

HTH