Many computer-aided control engineering packages have an option to export a model to C or C++ code. Often this option is used to solve a calculation-intensive problem in a quicker way, because the exported model does not suffer from the overhead from the package, but this code can also be ported to other environments. DUECA currently offers the possibility to include Simulink models converted with Simulink Coder.

This description is based on Matlab 2016b, with Simulink 8.11. Packages with Simulink 8.13 (2017b) and 9.0 (2018b) are also available. You need several items for these steps:

Of course a current Matlab license with the Simulink Coder enabled, to convert your Simulink model to c.
A package with the header files and compiled library code for any auxiliary code that your c-converted Simulink model might use. At TU Delft we use rpm or deb packages for this, the librtwxxx and librtwxxx-devel or librtwxxx-dev.
From librtw8_11 onwards the packages have a settings file, /usr/share/librtw8_11/dueca_rtw_config.m, that automatically adjusts to the right settings for code generation.

To generate code you need the following options:

Run dueca_rtw_config on your model. This sets the correct properties for the code generation.
Under "Solver", this selects a fixed step solver. You may specify a new step size. Note that the step size must match your DUECA update rate – if needed, you can also tweak the DUECA code to call your Simulink model several times per DUECA time step. Open the configuration parameters dialog (Simulation – Model configuration parameters).
Under "Code Generation", select ert.tlc as system target is selected, use language "C", and check for "Generate code only"
Under "Code Generation -- Comments", select all you want to see in your C code
Under "Code Generation -- Interface", make sure that "Reusable function" is selected for the packaging, and as well as "Use dynamic memory allocation" and "Terminate function required". Keep "Classic call interface" and "Single output/update function" unchecked. Don't check "MAT-file logging". For the option "Pass root-level IO as:", select "Structure reference".

Calling the Simulink Coder code

Generate a skeleton for your module by invoking the new-module script (use the right rtw version):

new-module rtw8_11

For calculation of the model response, you first need to set the input vector values. Look in doCalculation for the template code. This version of the generated code uses the Simulink Coder generated structs of IO, the commented code in the generated file should get you going, inspect the generated header file to see the structure of the inputs and outputs.

Linking

To link in the proper library, add the Simulink library as a component to DUECA, in the DCOMPONENTS variable in the main makefile, –rtw8_11 or –rtw8_13, –rtw9_0 or –rtw9_2 currently. This library already includes compiled code from some common generated files such as rt_nonfinite.c, rtGetNaN.c, rtGetInf.c files. These are generated automatically, but should not be included in the sources you compile with the DUECA module.