Weighing experiments in System Identification Toolbox

Hi everybody,

I'm using the System Identifcation Toolbox to identify nonlinear grey box models. The system has only one single input and one single output channel and there are 17 experiments, which should be reproduced by the nonlinear model.

The problem I have within this case is, that these different experiments all have different orders of magnitude both in the input as well as in the output channel. These leads to a result, which (because of the different residues) satisfies the experiments with high orders of magnitude more than the experiments with low orders of magnitude. Therefore my question: Is there a possibility in the System Identification Toolbox to weigh the different experiments, so the derived model satisfies all experiments in the same manner.

Because I'm working with nonlinear models a simple scaling of the input and output channles is not expedient.

Thank you very much in advance for your answers!