how to run a 3x3 matrix through a for loop and save all instances of the [3x3]

I have a [3x3] transformation matrix A = [ -sind(omega*t), cosd(omega*t), 0; cosd(omega*t), sind(omega*t) cosd(phi); cosd(omega*t), sind(omega*t), sin(phi)]

omega and phi are constants defined earlier. t is the vaiable that is being looped for. if I have 300 values for t, I would like to output 300 different [3x3] matricies, each one corresponding with a different value of t. for more information t is a value that is being read in from a data sheet.

[Adam Danz]

Is t a row or column vector?

[Tommy Pestolesi]

column

Answer 1

[Adam Danz]

A is a cell array the same size as t where A{n} is the matrix for t(n).

omega = 1; 
phi = 1; 
t = linspace(0,5,300);
A = arrayfun(@(t)[-sind(omega*t), cosd(omega*t), 0; 
    cosd(omega*t), sind(omega*t) cosd(phi); 
    cosd(omega*t), sind(omega*t), sin(phi)],t,'UniformOutput',false);

[Tommy Pestolesi]

Thank you that's just what I was looking for. Now that I have all of the entries of A my goal is to multiply all of the instances of A by a matrix that is made up of 3 column vectors. My matrix for the 3 column vectors can be refferred to as V. I want to multiply the first entry of A[3x3] by the first entry of V [3x1]so that I end up with as many [3X1] entries. Just trying to multiply the matrix by the vector.

[Adam Danz]

I'm not sure I follow. This line below multiplies each element of A by each column in V.

m is each single element of A.

y = cellfun(@(m)[m*V(:,1),m*V(:,2),m*V(:,3)],A,'UniformOutput',false);

Check out the documentation for cellfun() so you can understand what's going on here.

https://www.mathworks.com/help/matlab/ref/cellfun.html

[Tommy Pestolesi]

when I run your line of code I get inner matrix dimensions must agree. So I tried this line.

y = cellfun(@(m)[m*V'],A,'UniformOutput',false);

This returns a 300x1 cell for y. This 300x1 cell contains 300 3x300 doubles but I am trying to get 300 3x1 doubles. So I think I am very close to my final output, it may be an error in my matrix multiplication, thank you for your help and insight.

[Adam Danz]

The cellfun line I proposed works with my inputs (using Matlab r2019b). Unless you're running an older version of Matlab (prior to r2016b, I believe), when you run that line with my inputs (from my answer), it should work for you, too.

If it works with my inputs but doesn't work with your inputs, how do your inputs differ from mine? Also, what release of Matlab are you using?

[Tommy Pestolesi]

I am using Matlab version R2016b, my input for UVW is a 300x3 double. I'm confused as to why it doesnt work for me. Is it because m is a 3x3 and the columns are 300x1? But if that were the case your code should give the same error I get, hmmmmmmm. I'm going to troubleshoot some more and see what I can figure out.

[Adam Danz]

What's UVW? The input t in my answer is a 1x300 array.

[Tommy Pestolesi]

Sorry I changed the name of V to UVW, for me t is a 300x1 double that is read in data from a txt file.

[Adam Danz]

Is UVW 300x1 (as in your previous comment) or is it 300x3 (as in your comment before the previous one)?

[Tommy Pestolesi]

UVW is a 300x3

[Adam Danz]

From your question, "t is the vaiable that is being looped for. if I have 300 values for t".

In reality you have 900 values for t.

If I use these inputs

t = rand(300,3); 
A = arrayfun(@(t)[-sind(omega*t), cosd(omega*t), 0; 
    cosd(omega*t), sind(omega*t) cosd(phi); 
    cosd(omega*t), sind(omega*t), sin(phi)],t,'UniformOutput',false);

A is a 300x3 cell array. I'm guessing that's not what you want.

It's now unclear how to apply each element of t to the transformation matrix. The current code is processing each single value of t. I suppose you want to input rows of t. Is that correct?

[Adam Danz]

maybe this is what you're looking for

t = rand(300,3); 
A = arrayfun(@(i)[-sind(omega*t(i,1)), cosd(omega*t(i,1)), 0; 
    cosd(omega*t(i,2)), sind(omega*t(i,2)) cosd(phi);  
    cosd(omega*t(i,3)), sind(omega*t(i,3)), sin(phi)],1:size(t,1),'UniformOutput',false);

or perhaps this

t = rand(300,3); 
A = arrayfun(@(i)[-sind(omega*t(i,1)), cosd(omega*t(i,2)), 0; 
    cosd(omega*t(i,1)), sind(omega*t(i,2)) cosd(phi);  
    cosd(omega*t(i,1)), sind(omega*t(i,2)), sin(phi)],1:size(t,1),'UniformOutput',false);

But I haven't put too much thought into it, mainly because I gotta run. Think about how the values of t are being fed into the arrayfun and how you want to apply those value to your transformation matrix. I can check back later.