Look-Ahead / Previewing feature for Nonlinear Model Predictive Control

I would like to use an NMPC approach to control a nonlinear MIMO problem. In the real problem I have a very reliable prediction about the reference signal and I would like to use this knowledge in the form of a look-ahead approach.

In the NMPC example below, at the bottom of the page it says "This could be improved by providing the controller MPC with a look-ahead reference signal, which informs the controller of the expected reference variation over the prediction horizon."

https://de.mathworks.com/help/mpc/ug/nonlinear-model-predictive-control-of-exothermic-chemical-reactor.html

In the linear MPC's I found the corresponding feature under the "Improving Control Performance with Look-Ahead (Previewing)". Is there a similar feature for the nonlinear MPC?

Answer 1

[Alex Ander]

it is sufficient to specify a reference vector, Matlab then solves the optimization correctly with a preview behavior. At the beginning, I had always given only the current reference value and this is then apparently held constant over the complete prediction horizon.