120 CHAPTER 3. FUNCTIONAL DESCRIPTION OF Xµ
0.1 1 100.01 100
Magnitude
0.5
1
1.5
2
2.5
3
3.5
0
4
mu analysis of robust performance
Kmu
Kinf
3.7.4 Constructing Rational Perturbations
For simulation purposes it is useful to be able to construct a rational approximation to
the ∆ returned by the µ calculation. The approach is to choose a ∆ at a particular
frequency, for example the one where µ is at a maximum, and obtain a MIMO system
which has a frequency response (gain and phase) equal to ∆ at that frequency.
The function for this purpose is function mkpert. The syntax is given below.
pertsys = mkpert(Delta,blk,mubnds)
This function takes as arguments the variables Delta, blk,andmubnds. The meaning of
these is identical to the mu case. The frequency selected for the interpolation is that
where the lower bound (in mubnds) is maximum. Alternatively the user can use
keywords to specify a frequency at which to do the interpolation and specify the norm of
the resulting pertsys. pertsys will be an all-pass system.
Monte-Carlo simulation approaches require the ability to generate random perturbations
having the correct block structure. The function for this purpose is randpert and its
usage is illustrated below.
