For this example, set up your parameter estimation problem using the Parameter Estimator app and generate MATLAB® code from it. For more information on generating MATLAB code from the app, see Generate MATLAB Code for Parameter Estimation Problems (GUI). Alternatively, you can also set up your estimation problem at the command line.
Next, split the generated MATLAB code just before the estimation objective function is defined. This results in two files a run
function and a setup
function, as described in Parameter Tuning for Digital Twins.
In the setup
function, add the following lines of code at the end to configure the experiment and simulation test objects for deployment and save them to a MAT-file.
Experiment_out = prepareToDeploy(Experiement);
Simulator = createSimulator(Experiment_out);
Simulator = prepareToDeploy(Simulator,p);
save ObjectsToDeploy Experiment_out Simulator p
In the run
function, add the following lines of code at the beginning of the function to include the Simulink model in the compiled code and load the objects that were saved in the setup
function.
%#function simulink_model_name.slx
load ObjectsToDeploy Experiment_out Simulator p
Next, add the following lines to load the experiment data and update the experiment object. For this example, assume that the experiment data is contained in the first three columns of a Microsoft® Excel® spreadsheet file named fname.xls
.
d = xlsread(fname);
u = timeseries(d(:,3),d(:,1));
y = timeseries(d(:,2),d(:,1));
Experiment_out = updateIOData(Experiment_out,'simulink_model_name/param1',u);
Experiment_out = updateIOData(Experiment_out,'simulink_model_name/param2',y);
For a detailed example showing how to deploy your parameter estimation problem using the Simulink Compiler, see Parameter Tuning for Digital Twins.