Should I use multiple instances of one uncertain parameter to represent uncertainty in similar systems when using "robstab" or "robuststab"?

Should I use multiple instances of one uncertain parameter to represent uncertainty in similar systems when using "robstab" or "robuststab"?

Or would it help to have different uncertain parameters for things that are unique but grouped together? For example, I build an actuator model with an uncertain gain but then just replicate that three times for the different flight control surfaces. Would it be better or worse to have three different uncertain parameters for the flight controls or one that is replicated?

Answer 1

[MathWorks Support Team]

Unless the exact same actuator is used in these different flight surfaces, it is generally recommended to have independent uncertain parameters for each control surface. In addition to being more realistic, having more independent parameters with fewer repetitions makes it easier and faster to solve the uncertainty problem.