How to implement Hermite spline in Simulink

Good morning everyone.

Can i ask if it there is a way to implement the Hermite spline (pchip) or Bezier curves in simulink?

In particular, i am using a look-up table to interpolate a curve between some point, but the cubic and akima splines don't do a good job. I was thinking about using a linear interpolating method and then smoothen up the curves, but i can't find any blockset.

Thank you very much in advance.

Answer 1

[Bjorn Gustavsson]

There is a way to make blocks that wrap a matlab-function (matlab-functions-simulink_5), shouldn't that solve this task?

HTH

[MARCO LORENZI]

Oh ok, i see what you are telling me.

Can i ask if you have any idea on how to implement it in a look-up table?

[Bjorn Gustavsson]

I'd implement such a spline-interpolation-function in the simplest way like this:

function Yq = my_pchip(xq)
persisten X Y
if isempty(X) % think this works
    load lookuptable.mat X Y
end
Yq = interp1(X,Y,xq,'pchip');

You could look at generating the PP (piece-wise polynomial form of the lookup-table) instead and use ppval instead of interp1, to possibly speed up the function...

[MARCO LORENZI]

Hi, i am sorry to bother again but i have 1 last question.

So in the picture i attached i wanted to show what i wanted to do (substitute the lookup tables). In the two block is going to enter a number between 0 and 90 deg.

The output of the upper block says that if the input is <15 the output must be 1 and if >15 the output must be 0. The transition from 1 to 0 is made between 15 and 30deg.

The output of the lower block says that if the input is <15 the output must be 0 and if >15 the output must be 1 The transition from 1 to 0 is made between 15 and 30deg.

I plugged this code for the upper block

function output = low_angles_Hermite_approximation(phase)

x = length(phase);

y = [0 0 0 0 0 0 0 1 1];

xq = [0 2.5 5 7.5 10 13 15 30 90];

output = pchip(x,y,xq);

I plugged this code for the lower block

function output = low_angles_Hermite_approximation(phase)

x = length(phase);

y = [1 1 1 1 1 1 1 0 0];

xq = [0 2.5 5 7.5 10 13 15 30 90];

output = pchip(x,y,xq);

Matlab returns the error "The first two inputs must have at least two elements"

[MARCO LORENZI]

i don't think it is hard but i do not know how to do it. Basically, based on the input, that goes from 0 to 90, the first block does this:

The second one this:

And the output is y.

[Bjorn Gustavsson]

To me it seems as you swapped the x and xq in the function. It ought to be:

function output = low_angles_Hermite_approximation(phase)
xq = phase;
y = [0 0 0 0 0 0 0 1 1];
x = [0 2.5 5 7.5 10 13 15 30 90];
output = pchip(x,y,xq);

When you program and run into errors like these read the help and documentation carefully - and calmly, far to often it is easy to rush that part. Also be prepared to use the debug-facilities of matlab, I typically turn on debug-stop after an error:

>> dbstop if error

Then you get a command-line prompt at the line where the error occurred with command-line access to all variables in the function and it is possible to step up in the function call-stack.