1
.TH lqg 1 "April 1993" "Scilab Group" "Scilab Function"
12
: \fVsyslin\fR list (augmented plant) in state-space form
15
: 1x2 row vector = (number of measurements, number of inputs) (dimension of
16
the 2,2 part of \fVP\fR)
19
: \fVsyslin\fR list (controller)
21
\fVlqg\fR computes the linear optimal LQG (H2) controller for the
22
"augmented" plant \fVP=syslin('c',A,B,C,D)\fR (continuous time) or
23
\fVP=syslin('d',A,B,C,D)\fR (discrete time).
25
The function \fVlqg2stan\fR returns \fVP\fR and \fVr\fR given the
26
nominal plant, weighting terms and variances of noises.
28
\fVK\fR is given by the following ABCD matrices:
29
\fV[A+B*Kc+Kf*C+Kf*D*Kc,-Kf,Kc,0]\fR where \fVKc=lqr(P12)\fR
30
is the controller gain and \fVKf=lqe(P21)\fR is the filter gain.
31
See example in \fVlqg2stan\fR.
33
lqg2stan, lqr, lqe, h_inf, obscont