How to set output type manually in matlab coder

When I turn a simple matlab function that looks like y = wdn(x) into C code with matlab coder, the resulting C function looks like this:

void wdn(const float x[15400], emxArray_real_T *y)

How can I set y's type as float* ? I don't want to use emxArray_real_T because it is causing me problems when I try to use the code on a microcontroller.

I read somewhere that these structures are for dynamic memory allocation, however turning off dynamic memory allocation in matlab coder's settings did not help.

Answer 1

[Denis Gurchenkov]

Hi Alireza,

The "emxArray_real_T" type indicates that the code generator does not know what is the maximum number of elements in "y". That is, it thinks that the number of elements in y may change from run to run, and it could not figure out an upper bound for it.

You could help the code generator to understand that y is bounded by using one of the techniques here:

https://www.mathworks.com/help/coder/ug/variable-size-data-when-dynamic-memory-allocation-is-disabled.html

If that does not help, maybe can attach the example you have (and the command line arguments you use) so that someone could take a closer look.

You also mentioned that the emxarray type creates an issue on the microcontroller. Why is so? That type is just a plain C struct type, and there is a header file that defines it, all you need to do is to make that header file available to your C compiler.