how extract specific values from matrix and build separate arrays?

Hi there!

I have a function that generates a 4x4 matrix with different values at every iteration (100 matrices all up). I then want to get the 4th coulmn 1st row values of all these matrices in one array and the 4th column 2nd row in another array.

I could only think of a for loop to go through every matrix one at the time and extract the elemts required in each array but no success

Here is an example of what I am trying to acheive.

Any help would be much appreciated!

T(95) = 
    0.4982         0   -0.8671     1.495
    0.8671         0    0.4982     2.601
         0        -1         0         0
         0         0         0         1

T(96) = 
    0.4989         0   -0.8666     1.497
    0.8666         0    0.4989       2.6
         0        -1         0         0
         0         0         0         1

T(97) = 
    0.4995         0   -0.8663     1.498
    0.8663         0    0.4995     2.599
         0        -1         0         0
         0         0         0         1

T(98) = 
    0.4998         0   -0.8662     1.499
    0.8662         0    0.4998     2.598
         0        -1         0         0
         0         0         0         1

T(99) = 
    0.4999         0   -0.8661       1.5
    0.8661         0    0.4999     2.598
         0        -1         0         0
         0         0         0         1

T(100) = 
    0.5000         0   -0.8660       1.5
    0.8660         0    0.5000     2.598
         0        -1         0         0
         0         0         0         1
         
for i=1:100
    a1(i)=T(i)(4,1)
    a2(i)=T(i)(4,2)
end

         %resut:
a1=[......1.495 1.497 1.498 1.499 1.5 1.5]
a2=[[...... 2.601 2.6 2.599 2.598 2.598 2.598]

[Walter Roberson]

T(95) = 
    0.4982         0   -0.8671     1.495
    0.8671         0    0.4982     2.601
         0        -1         0         0
         0         0         0         1

Is that form, complete with the (95), being output to a text file? Or is that being output to the command window? Do you have access to the source code to be able to put the outputs into a cell array or a 3D array?

[Seba.V]

no that is out of the cooman window, and the function only give the matrix in that form.

Answer 1

[KALYAN ACHARJYA]

One Way: Indexing

%generates T in 3D 
T(:,:,1)= [0.4982         0   -0.8671     1.495
    0.8671         0    0.4982     2.601
         0        -1         0         0
         0         0         0         1]
T(:,:,2)= [0.4989         0   -0.8666     1.497
    0.8666         0    0.4989       2.6
         0        -1         0         0
         0         0         0         1]
T(:,:,3)=[0.4995         0   -0.8663     1.498
    0.8663         0    0.4995     2.599
         0        -1         0         0
         0         0         0         1]
T(:,:,4)= [0.4998         0   -0.8662     1.499
    0.8662         0    0.4998     2.598
         0        -1         0         0
         0         0         0         1]
T(:,:,5)=  [0.4999         0   -0.8661       1.5
    0.8661         0    0.4999     2.598
         0        -1         0         0
         0         0         0         1]
T(:,:,6)= [0.5000         0   -0.8660       1.5
    0.8660         0    0.5000     2.598
         0        -1         0         0
         0         0         0         1]

Then easy to extract the required data

>> whos T

Name Size Bytes Class Attributes

T 4x4x6 768 double

>> ar1=T(1,4,:)
ar1(:,:,1) =
    1.4950
ar1(:,:,2) =
    1.4970
ar1(:,:,3) =
    1.4980
ar1(:,:,4) =
    1.4990
ar1(:,:,5) =
    1.5000
ar1(:,:,6) =
    1.5000

Later you can go for 3D o 1D

[Seba.V]

is there a way to tranform the "normal" matrix into 3D? The function i am using only genrates the matrix as shown!

[KALYAN ACHARJYA]

This is not allow in MATLAB

T(95) = 
    0.4982         0   -0.8671     1.495
    0.8671         0    0.4982     2.601
         0        -1         0         0
         0         0         0         1

As Matrix are 2D, here you can assign in cell array or 3D. How you generates T, can you show that, or manual inputs?

[Seba.V]

I am using a free robotic toolbox (Peter Corke) function for forward kinematics called fkine()

[Walter Roberson]

The Peter Corke fkine() method returns an SE3 object. If you have that SE3 object, then you can use the R method to get a 3D array of rotations.

t = [0:.056:2]; 	% generate a time vector
q = jtraj(qz, qr, t); % compute the joint coordinate trajectory
T = p560.fkine(q);

Now

Rots = T.R;

Rots would be 3 x 3 x 36 where 36 is the number of poses (one for each t value in the above)

[Seba.V]

Amazing Walter...

Thank you!