MATLAB: Exported timeseries does not look the same as the imported one

interpolationsimulinktimeseriesworkspace

I created a time series in MATLAB using the "timeseries" function. If I import this time series to my model using a "From Workspace" block and then I try to export the same time series using a "To Workspace" block in a "Timeseries" save format, the two curves do not look the same when I plot them. Why is this happening?

Best Answer

The differences in the two curves are due to the interpolation methods of the time series being different. The "To workspace" block by default uses "zero order hold" interpolation, while the objects created using the "timeseries" function use linear interpolation by default. This can be resolved by setting the interpolation methods to be the same. For example, if "in_ts" represents the time series object created in MATLAB, the following would change the interpolation method to "zoh":

in_ts = setinterpmethod(in_ts,'zoh');

Related Solutions

MATLAB: Does “setVariable” in Simulink Test not update variables

In R2019b, the behavior of "setVariable" is modified such that it overrides only model variables (as opposed to MATLAB variables).

I attached a simpler model ("SimpleVars_R2019b.slx") to help illustrate this. This model has two "From Workspace" blocks, one set to "a" and one set to "evalin('base', 'b')".

If you call "findVars" on this model, only "a" is considered to be a model variable, even though both output their timeseries when run:

>> vars = Simulink.findVars('SimpleVars_R2019b')
vars = 
    VariableUsage with properties:
            Name: 'a'
          Source: 'base workspace'
      SourceType: 'base workspace'
           Users: {'SimpleVars_R2019b/From↵Workspace'}

The R2019b behavior provides stricter control of overrides and achieves parity with other Simulink tools. You will need to update the model to use variables which are recognized by Simulink.

Please update the "From Workspace" block variable parameter from "evalin('base', 'y_coordinate')" to just "y_coordinate". Then save the model and run the test file again.

You should observe that "setVariable" updates all of the variables now.

MATLAB: How can you view the function call event for a function call subsystem

In the Function-Call Subsystem, you can launch the block parameters of the "Trigger Port" block and check the "Show output port" option. You can then connect the output with a "Scope" or "To Workspace" block. In this way, every time the subsystem is called, the signal is logged. If you connect to a Scope, you can open the Scope setting dialog. In the "Style" tab, choosing a "Marker" will make the output more clear.

The Trigger block will output a value as:

1 for a signal that causes a rising trigger
-1 for a signal that causes a falling trigger
2 for a function-call trigger
0 otherwise

Best Answer

Related Solutions

MATLAB: Does “setVariable” in Simulink Test not update variables

MATLAB: How can you view the function call event for a function call subsystem

Related Question