ui_observer

unknown input observer

Calling Sequence

[UIobs,J,N]=ui_observer(Sys,reject,C1,D1)
[UIobs,J,N]=ui_observer(Sys,reject,C1,D1,flag,alfa,beta)

Arguments

:Sys syslin list containing the matrices (A,B,C2,D2). : :reject integer vector, indices of inputs of Sys which are

unknown.

: :C1 real matrix : :D1 real matrix. C1 and D1 have the same number of rows. : :flag string ‘ge’ or ‘st’ (default) or ‘pp’. : :alfa real or complex vector (loc. of closed loop poles) : :beta real or complex vector (loc. of closed loop poles) :

Description

Unknown input observer.

Sys: (w,u) –> y is a (A,B,C2,D2) syslin linear system with two inputs w and u, w being the unknown input. The matrices B and D2 of Sys are (implicitly) partitioned as: B=[B1,B2] and D2=[D21,D22] with B1=B(:,reject) and D21=D2(:,reject) where reject = indices of unknown inputs. The matrices C1 and D1 define z = C1 x + D1 (w,u), the to-be-estimated output.

The matrix D1 is (implicitly) partitioned as D1=[D11,D12] with D11=D(:,reject)

The data (Sys, reject,C1, D1) define a 2-input 2-output system:

xdot =  A x + B1  w + B2  u
   z = C1 x + D11 w + D12 u
   y = C2 x + D21 w + D22 u

An observer (u,y) –> zhat is looked for the output z.

flag=’ge’ no stability constraints flag=’st’ stable observer (default) flag=’pp’ observer with pole placement alfa,beta = desired location of closed loop poles (default -1, -2) J=y-output to x-state injection. N=y-output to z-estimated output injection.

UIobs = linear system (u,y) –> zhat such that: The transfer function: (w,u) –> z equals the composed transfer function: [0,I; UIobs Sys] (w,u) —–> (u,y) —–> zhat i.e. transfer function of system {A,B,C1,D1} equals transfer function UIobs*[0,I; Sys]

Stability (resp. pole placement) requires detectability (resp. observability) of (A,C2).

Examples

A=`diag`_([3,-3,7,4,-4,8]);
B=[`eye`_(3,3);`zeros`_(3,3)];
C=[0,0,1,2,3,4;0,0,0,0,0,1];
D=[1,2,3;0,0,0];
`rand`_('seed',0);w=`ss2ss`_(`syslin`_('c',A,B,C,D),`rand`_(6,6));
[A,B,C,D]=`abcd`_(w);
B=[B,`matrix`_(1:18,6,3)];D=[D,`matrix`_(-(1:6),2,3)];
reject=1:3;
Sys=`syslin`_('c',A,B,C,D);
N1=[-2,-3];C1=-N1*C;D1=-N1*D;
nw=`length`_(reject);nu=`size`_(Sys('B'),2)-nw;
ny=`size`_(Sys('C'),1);nz=`size`_(C1,1);
[UIobs,J,N]=ui_observer(Sys,reject,C1,D1);

W=[`zeros`_(nu,nw),`eye`_(nu,nu);Sys];UIobsW=UIobs*W;
//(w,u) --> z=UIobs*[0,I;Sys](w,u)
`clean`_(`ss2tf`_(UIobsW));
wu_to_z=`syslin`_('c',A,B,C1,D1);`clean`_(`ss2tf`_(wu_to_z));
`clean`_(`ss2tf`_(wu_to_z)-`ss2tf`_(UIobsW),1.d-7)
/////2nd example//////
nx=2;ny=3;nwu=2;Sys=`ssrand`_(ny,nwu,nx);
C1=`rand`_(1,nx);D1=[0,1];
UIobs=ui_observer(Sys,1,C1,D1);

See Also

  • cainv Dual of abinv
  • ddp disturbance decoupling
  • abinv AB invariant subspace

Table Of Contents

This Page