how to obtain 1-step prediction with narx?

Assume that I obtained a narx model using system identification toolbox based on the regressors y(k) and u(k). I.e.,

y(k+1)=f(y(k), u(k)).

How can I calculate y(k+1) given for example that y(k)=1, u(k)=-1? I.e., how can I calculate the one-step prediction y(k+1) GIVEN y(k) and u(k)?

[Gabriel Theberge]

I have the EXACT same question, and it seems that I'm not alone. There is a lot of people who experiment the same problem, and this question is unsolved everywhere I saw it on internet. I think the Matlab documentation is unclear about that. If someone has the answer to this question (or a link to the answer), I will be very grateful if you share it. Thank you very much.