I want to proof that my measured signal is identical with my actual signal.

[medum]

I receive measurements that run simultaneously on a chart in real-time. I want to check these measurement points to see if they correlate with my actual signal. That means, I collect measurements and construct my signal from them, and then see if they correlate with my actual signal with a certain deviation. How would this be done in LabVIEW.

I am not sure what is an clever solution.  

[jamiva]

Google "cross-correlation"

[altenbach]

A chart is a lossy UI device with a fixed history where data beyond the history length is irreversibly lost. You probably want to compare the incoming scalar(?) data with a lookup based on elapsed time and other instrumental parameters (temperature, pressure, etc.), but currently your question is too vague to really offer a suitable solution. You need to be much more specific.

Maybe a simple range checking would be sufficient (i.e. Is new signal in rage of old signal +/- allowed deviation?)

We can typically give much more specific help if you would attach a simple Vi containing typical data. For example we don't know the datatype of your "signal".

• Dynamic data?
• Waveform data? If so, do both have the same t0, dt?
• One scalar point at a time?
• 1D integer array?
• 1D DBL array?
• 1D complex array?
• etc.etc

What are the comparison parameters for "equal"?

Are the signals allowed to have different DC offsets but must have the same shape?

Should they have similar noise? What is the noise (Gaussian? shot?)

Do both have the same number of points?

Can the data be parameterized with a simple model (polynomial, exponential, etc.) as a function of time?

So. Many. Questions.

[medum]

Thank u for your respond.

I am measuring the data in real time -> One scalar at time. 

To see if it's a sawtooth wave, and I'd like to randomly inspect it over a period of 10 seconds.

My aim is to reconstrucate this signal und check if this signal similar with an other signal. It is allowed to have little error margin. 

[altenbach]

Sorry, can't look at your code at the moment because your version is too new. Consider "save for previous" before attaching (2020 or below).

[medum]

Sorry...

Here is the version V20.

[Bob_Schor]

There are some puzzling things you mentioned in your earlier posts.

• What do you mean by "my measured signal"?  Are you sampling a (voltage) signal using LabVIEW and a digital sampling device (an Analog-to-digital converter, such as a USB NI-DAQ device) or other measuring instrument?
• What is the source of your "actual" signal?  Are you using a signal generator whose behavior you trust?  
• The code snippet you attached shows LabVIEW generating a Waveform by simulating hardware and "pretending" that this is a "real" (voltage) signal.
• If you really want to "measure" a signal, you do not want to "randomly inspect it over a period of 10 seconds" -- you want to "sample" it at a frequency at least 10 times higher than the highest frequency you are interested in seeing (so if you are generating a 100 Hz sawtooth wave, you'd sample at 1000 Hz or higher).

Bob Schor

[altenbach]

---

@medum wrote:

Sorry...

Here is the version V20.

---

Thanks, It typically helps to have reasonable default values for all controls so we can reproduce whatever you are doing. I am sure that a frequency and amplitude of zero are not reasonable. Do you know how to define defaults?

According to the configuration of the express VI, you are generating 100 point at a time, and connecting a scalar indicator to the dreaded dynamic data does not change that fact. You also only have one signal, nothing to compare. This VI has very little to do with your original description.

I recommend to start from scratch and explain in detail what you are actually trying to do. 

[altenbach]

See if this simple simulation can give you some ideas.

[medum]

Thank u very much.

Could u explain me, the block diagram.

And why I need Gaussian Noide SD?