Cannot compute Riccati solution for LQI controller

70 次查看（过去 30 天）
Leonardo Costa
Leonardo Costa 2021-4-7
评论： Leonardo Costa 2021-4-8
Hi to everybody!
This is my code right now:
u=0.5;
v=0;
omega=0;
m=50;
psi=0;
c_d_Long=0.5;
c_dr=0.6;
c_d_Lat=1;
rho=1000;
L=1.2;
W=0.275;
H=0.416;
I=38;
A =[0, 0, - v*cos(psi) - u*sin(psi), cos(psi), -sin(psi), 0;
0, 0, u*cos(psi) - v*sin(psi), sin(psi), cos(psi), 0;
0, 0, 0, 0, 0, 1;
0, 0, 0, -((3*W*c_d_Long*rho*u^2*1)/10 + (3*W*c_d_Long*rho*u*sign(u))/10)/m, omega, v;
0, 0, 0, -omega, -((3*L*c_d_Lat*rho*v^2*1)/10 + (3*L*c_d_Lat*rho*v*sign(v))/10)/m, -u;
0, 0, 0, 0, 0, - (300*L*c_dr*omega^2*rho*1)/38347 - (300*L*c_dr*omega*rho*sign(omega))/38347];
B =[ 0, 0, 0, 0, 0, 0, 0, 0;
0, 0, 0, 0, 0, 0, 0, 0;
0, 0, 0, 0, 0, 0, 0, 0;
1/m, 1/m, 1/m, 1/m, 0, 0, 0, 0;
0, 0, 0, 0, 1/m, 1/m, 1/m, 1/m;
825/76694, -825/76694, 825/76694, -825/76694, 325/38347, 325/38347, -325/38347, -325/38347];
C=[1 0 0 0 0 0;
0 1 0 0 0 0;
0 0 1 0 0 0;
0 0 0 1 0 0;
0 0 0 0 0 0;
0 0 0 0 0 1];
SYS=ss(A,B,C,0);
Co = ctrb(SYS);
C=rank(Co);
Ob=obsv(SYS);
O=rank(Ob);
Q=eye(12);
R=eye(8);
klqi=lqi(SYS,Q,R)
As you can see I have a 6-dimension state with 8 input and 6 output. When i Run this code i have the following error:
Cannot compute a stabilizing LQR gain (the Riccati solution S and gain matrix K are infinite).
This could be because:
* A has unstable modes that are not controllable through B,
* Q,R,N values are too large,
* [Q N;N' R] is indefinite,
* The E matrix in the state equation is singular.
I had the same issue in the past but every time the problem was the controllabilty of the state but in this case I haven't this problem (rank of the controllabilty matrix= 6)
Hope somebody can help me.
Thank you in advance.

请先登录，再回答此问题。

采纳的回答

Paul
Paul 2021-4-8
编辑：Paul 2021-4-8
I'm not sure it's relevant, but should the fifth row of SYS.c be all zeros?
I suspect the issue is with stabilizability of the augmented system. Note that the code as posted is checking controllability of the plant, but the LQR solution requires stabilizability of the augmented plant, i.e., the plant combined with the integral control. I think the augmented system (not incuding the reference input) can be formed as
sysaugmented = series(SYS,ss(zeros(6),eye(6),eye(6),0)); % with SYS.c = eye(6)
Then the rank of the controllability matrix is
>> rank(ctrb(sysaugmented))
ans =
9
So the augmented system is not controllable. The eigenvalues are:
>> eig(sysaugmented)
ans =
0
0
0
0
0
0
0
0
0
-0.6188
0
0
So at least two of the uncontrollable modes are on the imaginary axis, which (I think) means the augmented system is not stabilizable, which violates one of the conditions for the solution to exist as stated here:
doc lqi
  3 个评论
显示 1更早的评论
Paul
Paul 2021-4-8
编辑：Paul 2021-4-8
Yes, you can augment the plant manually with the integrator and then use the lqr() command to compute the feedback gains,subject to the standard consraints for LQR solutions.

请先登录，再进行评论。

更多回答（0 个）

类别

在 Help Center 和 File Exchange 中查找有关 State-Space Control Design 的更多信息

标签

产品


版本

R2020b

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by