From what I understand, you want to identify 1-minute-long segments of your data which represents steady state. Are “RQ”, “VE”, and “VO2” columns of your table? Also, what is the time step between each point in the data? Assuming each of these quantities can be accessed as vectors (“RQ”, “VE”, and “VO2”) and your time step is 10 seconds, you can use the following code:
time_step = 10; % seconds
num_rows = 37;
rows_per_minute = 60 / time_step;
percent_change = @(x) 100 * abs((x - x(1)) / x(1));
steady_data_indices = [];
% Set up dummy data for RQ, VE, and VO2
a = 90; b = 100;
RQ = (b-a) .* rand(1, num_rows)' + a;
VE = (b-a) .* rand(1, num_rows)' + a;
VO2 = (b-a) .* rand(1, num_rows)' + a;
for i = 1:(num_rows - rows_per_minute) % sliding window
RQ_percent_change = percent_change(RQ(i:i+rows_per_minute));
VE_percent_change = percent_change(VE(i:i+rows_per_minute));
VO2_percent_change = percent_change(VO2(i:i+rows_per_minute));
if all(RQ_percent_change < 5) && all(VE_percent_change < 10) && all(VO2_percent_change < 10)
steady_data_indices = [steady_data_indices i];
end
end
This code takes a 1-minute-long “window” and slides it through the data, checking the percent deviation of “RQ”, “VE” and “VO2” from the first value of the window to make sure they are within the bounds you specified. After the for loop completes, “steady_data_indices” will contain all of the starting indices where there is a 1-minute-long segment of steady state data.
Depending on how you define percent change, you may need to use a different function handle for “percent_change”, but hopefully this helps you get started.
