MATLAB: Does Simulink throw an error about variable size matrix from the MATLAB Function block

functionMATLABsimulinksize;variable

Why do I received the following error in Simulink?

Data 'x' is inferred as a variable size matrix, while its properties in the Model Explorer specify its size as inherited or fixed. Please check the 'Variable Size' check box and specify the upper bounds in the size field.

Component:MATLAB Function | Category:Coder error

Simulink cannot determine sizes and/or types of the outputs for block 'MATLAB Function' due to errors in the block body, or limitations of the underlying analysis. The errors might be inaccurate. Fix the indicated errors, or explicitly specify sizes and/or types for all block outputs.

Component:MATLAB Function | Category:Coder errorOpen

Simulink cannot determine sizes and/or types of the outputs for block 'MATLAB Function' due to errors in the block body, or limitations of the underlying analysis. The errors might be inaccurate. Fix the indicated errors, or explicitly specify sizes and/or types for all block outputs.

Component:Simulink | Category:Model error

Error occurred in 'example/MATLAB Function'.

Component:Simulink | Category:Model error

My MATLAB function only has these few lines of code:

function [x] = fun(A)
y = max(A);
x = find(A==y,1);

Best Answer

This error occurs because the "find" function always returns a variable-length vector.

If you do not want to change the properties of the variable "x" in MATLAB Function to be variable sized as suggested in the error message, you can change the MATLAB Function to not use "find".

For example,

function [x] = fun(A) 
[y, x] = max(A);

Related Solutions

MATLAB: Are the MATLAB Function block arguments being inferred as variable size

In this particular case, variables X and Y were defined as Tunable parameters within the MATLAB Function block. This leads to Simulink interpreting the array as variable sized because X and Y could theoretically change during the simulation, even though their values clearly do not change here.

To resolve this, there are two potential solutions:

Set X and Y to be non-tunable using the 'Edit Data' option from within the MATLAB Function Block.
Set 'out' to variable-size as suggested by the message. In most cases this will resolve the warning straight away, but there may be other limitations that cause warnings/errors elsewhere in the model, so in general this should be recommended second.

Other users have encountered the same thing here:

https://www.mathworks.com/matlabcentral/answers/79239-error-due-to-variable-size-data-in-matlab-function-block

MATLAB: What is the cause of “In numerictype(s) s must be a scalar constant or a fi” error

The error is caused by the fact that built-in integers, like the int16 constant in the example model you kindly provided, are not treated as fixed point object.

There is two possible workarounds in this case.

1. You can set the treatment of inherited Simulink signal types to 'Fixed-point & integer' in the Ports and Data Manager, that is available from the toolbar when editing the MATLAB Fucntion Block. This is shown in the image below

2. You can ensure the casting of the input data yourself, so that integers will always be converted to fi objects. The following code segment achieves that and could be used in the MATLAB Function Block:

>> if isinteger(u_in) 
>>     u = fi(u_in); 
>> else 
>>     u = u_in;
>> end

Please note that the input argument has been changed to 'u_in' in order to allow for proper casting.

Related Question