I want to separate each gait cycle and keep each cycle in a row of a table.

I have an excel file (attached here) having a column of data that has data points of several consecutive vertical GRF cycles. I want to separate each gait cycle and keep each cycle in a row of a table.How I can do it using MATLAB?

Thanks in advance.

Answer 1

[Eric Sofen]

If you separate each cycle into a separat row, the rows will likely have data of different lengths, which then gets hard to work with because you'll need to store them in cell arrays. As a starting point, here's what I'd do:

t = readtable("h1l_130710_1_110.xlsx");
t.Min = islocalmin(t.Fz,"MinProminence",500); % this works pretty well. 
% % I recommend using the "Find Local Extrema" live task to play around with peak-finding parameters interactively.
t.Cycle = cumsum(t.Min);
t.Min = [] % probably don't need this any more

% now you have Cycle as a grouping variable and you can use things like
% groupsummary or varfun to do calculations on each gait cycle.

%If you really want the data as rows...
t = varfun(@(x) {x'}, t, "GroupingVariables","Cycle")

[Mohammad Ahmad]

thanks a lot. I found it completely helpful. one more thing i need to know, if I want to overlap all rows using plot as shown in the attached figure , what to do then?

Thanks again.

[Eric Sofen]

Given that the gaits are different lengths, you need to normalize the spacing of the data. Starting from above the last line of code.

t1 = varfun(@(x) (1:numel(x))'/numel(x), t, "GroupingVariables","Cycle");
t1.Properties.VariableNames("Fun_Fz") = "GaitFraction";
t = [t,t1(:,3)];

% you can either use another grouped varfun for plotting, or at this point
% a for-loop is easier to understand.
hold on;
for i = 1:max(t.Cycle)
    gaitT = t(t.Cycle == i, :);
    plot(gaitT.GaitFraction, gaitT.Fz)
end

There's some fine-tuning to be done, but this should get you going the right direction.

[Mohammad Ahmad]

thank you for your response. Here the variable Fun_Fz is unrecognized. could you please address it? I tried several times but could'nt find any way.

[Mohammad Ahmad]

please address this issue