Grouping using for loops (signal processing)

I am working with a signal data that consist of consecutive dips as shown below. I am only interested in the portion of the signal that lies below a certain point d (the red line). I am trying to write a code which sorts the contents of each dip into one separate group.

And here is the grouping that I need:

For instance, the following code is one of my attempts which didn't work. It generates 310 groups instead of the desired 12 groups.

k=0; % Group number
 
for i = 1 : length(signal)
    if signal(i) < d
        k=k+1;
        while signal(i) < d
            NewSignal(i, k) = signal(i);
            i = i + 1; 
        end
    end
end

Any explanations or suggestions would be greatly appreciated.

P. S. I've included an attachement of my data which can be accessed as follows:

M = csvread('DS0007.csv');
time = M(:,1)*1.00e-03;
waveform = M(:,2)*2.00e-01;
waveform = sgolayfilt(waveform,9,21);

[dpb]

Attach a data file so folks have something to work with.

My suggestion/starting point would be to use findpeaks to locate each peak and then pick the data from those locations either side the peak to the threshold.

[Sordin]

Hi dpb, I've attached my sample data. I did try findpeaks(...), but functions like these find more than one peak per dip. To follow your suggestion I would need something that only pinpoints the absolute lowest point of each dip. Besides, I would prefer not to use the signal processing toolbox because some of the computers I am working with are not equiped with that.

[dpb]

You're already using it...

>> which sgolayfilt
C:\ML_R2017\toolbox\signal\signal\sgolayfilt.m
>> 

But, it appears the signal is noise-free enough that it shouldn't be hard to find the locations with a much larger difference between ordinal positions than the next point once you've separated out those less than the threshold.

Use find to return the positions from the logical vector

ix=find(waveform<d);
peakarray=waveform(ix);
dx=diff(ix);