Hi Amir

i know these ways that you suggested and many thank, but my problem is that I have a Continuous model that feeds with just Continuous signals. I can't find a true way to feed it with a Continuous signal. as I have a discrete samples.

my Matlab function is a Neural network that interpolate entered data exactly but the output of this block is discontinuous and I can't feed it to my continuous dynamic...!

I use S.Madgwick Filter with aim of magnetic distorsion ,bias , drift, reduction.

Amir Shakouri

Hello

There is a block in simulink 'the repeating sequence bloc' that converts a vector to a periodic signal. You can search for that and use it for the frequency that you want.

Cheers

Salah Eddine Ghamri thanks a lot. but the repeating sequence block doesn't have any input port. and i can't feed it from any where like workspace!

@Ahmad Abbasian

Are you sure about that ?! I don't remember quite well it's usage but once I have used it solely for your purpose.

Anyway a signal in simulink is a time series. You can copy it's structure from "toworkspace" then build your signal using that and some kind of interpolation.

Cheers

Use an Integrator (a filter)

If you have a vector representing a time-series sequence, then most basic block you are describing is the 'Zero Order Hold' (ZOH) block, which extends a sample into continuous time (more basic than an integrator). However, there are many answers in this thread that are worth exploring. It really depends what you are doing with the data and what information you have about its behaviour between samples (if any).

The ZOH block is a simple in/out block (I think there's an enable for clock/reset as well is some versions of Simulink).

If it's a vector of data that you want to curve-fit, then that is best done with an S function block and some code that calls 'cscvn' and sends it to the output port.