How to implement regressionGP class in Simulink

Hello All,

I am trying to use a Gaussian process regression model in Simulink. My setup is very basic, yet I cannot get it to work. The function block simply loads the model, then predicts the output given the new observations. When running the contents of the function outside of Simulink it runs with no errors, but when I run from within Simulink, there are multiple errors from within the 'predict' methods.

I've attached the .m file and .slx file, if anyone has any advice that would be great.

Mathworks has an example on executing classification models within Simulink, but I don't see anything on GPR models.

Thanks

Answer 1

[Salman Ahmed]

Hi Holden,

From my understanding, you would like to replicate the functionality of your MATLAB file in Simulink. The functionality is replicated correctly. The error you are getting in Simulink is due to incorrect signal specification. Add a signal specification block with the size of the inputs specified in dimensions [1 11] (as shown), your errors will resolve.