Hi.
I'm using Labview 8.6 combined
with Matlab between Matlab Script Node and I have some problems. Im using
Matlab R2008a. In Matlab i’m using the ident toolbox to identify some processes
given a data text archive.
I have a function that
depends on another functions that doesn't works fine when it’s called from
Matlab Script Node.
I have test the function on
Matlab and it shows me the proper results. But when I put it on Labview with
Matlab script Node (I’ve also probed to put it on Mathscript node and it doesn’t
work either) doesn’t give me the correct values (the same values that I have
obtained when I’ve call the function from Matlab).
I think that the problem is
in my code, so that I will give you all the functions I’ve used for accomplish
my purposes and the used VI.
(I’ve only revised* the
functions made by me, and not the rest of them, I suppose that they don’t have
characteristics that Matlab don’t give support to )
*that there aren’t any
characteristics of Matlab language that Matlab Script Node don’t support. (I
read on NI forums that Labview (actually Matlab Script Node) doesn’t use Matlab
structures correctly when you put them in the form “variable.field” (using the “.”
Operator) and because of this fact, i have modified the syntaxis of all
the structures that i put before; moreover: i’ve read that it also doesn’t
works fine with arrays of 3D or more ).
To resume, I will be so
happy if anybody wants to read my code for searching any incompatibility
between certain expressions and Matlab Script Node.
In order to do the work easier, I copy the list of variables returned
and arguments that are necessary to send to the main function
“AsignacionPolosssVER2”. Moreover, I put it, certain values for the inputs and
the CORRECT values of outputs of the function for these certain values, in
order to set a reference to distinguish if the VI runs fine or not.
Archives made by me:
-scriptaparte.vi
-saturacion.m
- cal_procesodiscreto.m
- cal_procesocontinuo2VER2.m
- cal_procesocontinuo2VER2.m
NOTE: change the path given
to the variables “pa” and “ruta_ar” (in “scriptaparte.vi”) “D:\Documents and
Settings\Administrador\Escritorio\scriptnode2\datos7p75.txt” (path + name of
data archive) and “D:\Documents and
Settings\Administrador\Escritorio\scriptnode2” (path) from the correct ones in
your case.
Main function prototype:
function
[numreg,denreg,Tmuestreo,status,num_gs,den_gs,delay,numd,dend,bestPer,bestMod,FITf,Modelable,sat]=AsignacionPolossVER2(metodo,espd,ptodis,espe,CancelacionCeros,ruta,FITumbral,n)
Charging certain values:
[numreg,denreg,Tmuestreo,status,num_gs,den_gs,delay,numd,dend,bestPer,bestMod,FITf,Modelable,sat]=AsignacionPolossVER2(0,[0
3 3],[-3 5],1,'s','I:\LabVIEW 8.6\examples\scriptnode2\datos7p75.txt',3,0)
You only have to copy this
on Matlab to view the results. However i put it the correct values of outputs
for this inputs:
numreg =[309.4818 -305.4845]
denreg =[1.0000 1.5445] Tmuestreo=0.015 status=1
num_gs = 0 -0.3310
0.7064
den_gs = 1.0000
0.2858 0 delay=0.45
numd=-0.0049 0.0050
dend= 1.0000
-1.9957 0.9957
bestPer = 91.2001
bestMod ={'P1IZD';}
FITf =91.2001
Modelable =1
sat =
1
526
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
My opinion: I think that the the program is
not well interpreted by Matlab Script Node because when the VI prints all the
outputs for all the models (output to the instruction compare(model1,...modeln),
the correct (the model which has the best fit -percentage-) model has a
different percentage of similarity to the data (in both cases i’m using the
same .txt to load the data). Some output variables are right (ie. Tmuestreo).
I give you the folder where
i’m working.
Sorry, but my english is not so good.
(If anybody speaks Spanish well, i accept to be replied on this language)
Thanks.
Oscar.
