From: omri374 Mendels on
Hello,

I'm trying to construct an interpolation using hermite polynomial (PCHIP).

this is a piece of my code:
%Xs is a two dimensional array of n rows and 2 columns: x and y.
%step is a parameter that has keeps the number of points along the curve.


x = Xs(:,1);
y = Xs(:,2);
t = Xs(1,1):step:Xs(n,1);
p = pchip(x,y,t);
arr = [t; p]';


I do get an interpolation, but the distances (euclidean) between points are not constant. I tried using another form of pchip:
pp = pchip(x,y) that returns a structure with coefficients but still haven't figured out how to use this in order to build the curve I need.

Any help will be appreciated!
Thanks,
Omri.
From: Bruno Luong on
You might check out John's FEX submission:
http://www.mathworks.com/matlabcentral/fileexchange/27096-interparc

Bruno
From: omri374 on
"Bruno Luong" <b.luong(a)fogale.findmycountry> wrote in message <hrjd4j$lal$1(a)fred.mathworks.com>...
> You might check out John's FEX submission:
> http://www.mathworks.com/matlabcentral/fileexchange/27096-interparc
>
> Bruno

Thanks! it helps me a lot!
From: John D'Errico on
"omri374 " <omrimendels(a)hotmail.com> wrote in message <hrjeso$dtj$1(a)fred.mathworks.com>...
> "Bruno Luong" <b.luong(a)fogale.findmycountry> wrote in message <hrjd4j$lal$1(a)fred.mathworks.com>...
> > You might check out John's FEX submission:
> > http://www.mathworks.com/matlabcentral/fileexchange/27096-interparc
> >
> > Bruno
>
> Thanks! it helps me a lot!

I must smile here, because I wrote that more for fun
than expecting anyone to want to use it. It just
seemed like a pretty application of an ode solver.

An interesting question is if I can force the system I
form to be a stiff one, thereby making ode45 a poor
choice of solver. I believe this to be true, so while
ode45 is indeed the best solver for this problem most
of the time, I should offer an alternative choice of
ode solvers.

One more thing to add to the list.

John