Error using horzcat Dimensions of arrays being concatenated are not consistent.

There is an error in the line with bold letters:

clear
Ad=[0.941 0.192; -0.576 0.903]; 
Bd=[.020; 0.192]; Gd=[.020; 0.192]; Sw=0.5; 
Dx=[1 0;0 1;0 0]; Du=[0;0;1]; 
zmin=[-0.4; -1e6; -3]; zmax=[0.4; 1e6; 3]; 
nx=2;nu=1;nw=1;nz=3; 
Q=[1 0;0 1]; R=0.5; S=[0; 0];
P=dare(Ad,Bd,Q,R); 
L=inv(R+Bd'*P*Bd)*Bd'*P*Ad; Qf=P;
mpc=1; % set to zero for unconstrained LQOC
delay=0; % set to zero no delay
if(mpc == 1)
    N=10; %horizon size
  Mu=zeros(nu*N);Mx=zeros(nx*N);Mux=zeros(nu*N,nx*N);
    for i=1:N-1
Mu(1+(i-1)*nu:nu+(i-1)*nu,1+(i-1)*nu:nu+(i-1)*nu)=R; 
Mx(1+(i-1)*nx:nx+(i-1)*nx,1+(i-1)*nx:nx+(i-1)*nx)=Q; 
Mux(1+i*nu:nu+i*nu,1+(i-1)*nx:nx+(i-1)*nx)=S';
    end
Mu(1+(N-1)*nu:nu+(N-1)*nu,1+(N-1)*nu:nu+(N-1)*nu)=R;
Mx(1+(N-1)*nx:nx+(N-1)*nx,1+(N-1)*nx:nx+(N-1)*nx)=Qf; 
M=[Mu Mux;Mux' Mx]; m=zeros((nu+nx)*N,1);
% Add Slack variables
    M=[M zeros(N*(nx+nu),nz);zeros(nz,N*(nx+nu)) 0.0*eye(nz)]
    m=[m; 1e6*[1; 1; 10]];
   A0u=zeros(nx*N,nu*N);A0x=zeros(nx*N,nx*N);
    A0u(1:nx,1:nu)=Bd; A0x(1:nx,1:nx)=-eye(nx);
    for i=2:N
        A0u(1+(i-1)*nx:nx+(i-1)*nx,1+(i-1)*nu:nu+(i-1)*nu)=Bd;
        A0x(1+(i-1)*nx:nx+(i-1)*nx,1+(i-1)*nx:nx+(i-1)*nx)=-eye(nx);
        A0x(1+(i-1)*nx:nx+(i-1)*nx,1+(i-2)*nx:nx+(i-2)*nx)=Ad;
end
    A0=[A0u A0x];        b0=zeros(nx*N,1);
    A0=[A0 zeros(nx*N,nz)];  % Add Slack variables
 
    A1u=zeros(nz*N,nu*N); A1x=zeros(nz*N,nx*N); A1u(1:nz,1:nu)=Du;
    b1max=zeros(nz*N,1); b1min=zeros(nz*N,1);
    A1slack=zeros(nz*N,nz); % Add Slack variables
    A1slack(1:nz,1:nz)=-eye(nz); % Add Slack variables
    for i=2:N
        A1u(1+(i-1)*nz:nz+(i-1)*nz,1+(i-1)*nu:nu+(i-1)*nu)=Du;
        A1x(1+(i-1)*nz:nz+(i-1)*nz,1+(i-2)*nx:nx+(i-2)*nx)=Dx;
        A1slack(1+(i-1)*nz:nz+(i-1)*nz,1:nz)=-eye(nz);
        b1max(1+(i-1)*nz:nz+(i-1)*nz,1)=zmax;
        b1min(1+(i-1)*nz:nz+(i-1)*nz,1)=-zmin;
end
     A1=[A1u A1x;-A1u -A1x];
     % Add Slack variables
     A1=[A1 [A1slack; A1slack];zeros(nz,N*(nu+nx)) -eye(nz)];
end

Error using horzcat

Dimensions of arrays being concatenated are

not consistent.

Error in Untitled10 (line 26)

M=[M zeros(N*(nx+nu),nz);zeros(nz,N*(nx+nu)) 0.0*eye(nz)];

[KSSV]

It is working fine in my case.....

[Harleen Kaur]

Did you just copy word to word?

Also, could you just try solving only '[M zeros(N*(nx+nu),nz);zeros(nz,N*(nx+nu)) 0.0*eye(nz)]' in the command window and see if it's working? I guess as a whole it's just ignoring instead of solving.

Actually this is an example of 2*2 matrix and it's working fine, but it's not working in 6*6 matrix.

TIA

[KSSV]

M=[M zeros(N*(nx+nu),nz);zeros(nz,N*(nx+nu)) 0.0*eye(nz)]

It is working fine while code is run. When you run it again onthe command window, obviously it will throw error, as the size of M is now modified.

[Harleen Kaur]

I am so sorry, but just 1 last thing. Could you see if it's working with these matrices? It's not working for me for 6*6 matrix. TIA

Ad=[0.478969081440410,-22.9347831060310,0,0.137256358561114,0.206216068991871,-1.92578298257640;-3.05094506276960e-05,0.0355133168622329,0,-0.000177637354131665,-6.76055868392685e-06,0.0265441846751822;3.05094506276902e-05,0.743287466209174,0.778800783071405,0.000177637354131653,6.76055868391955e-06,0.154257043717761;0.143460556023548,-150.655973086421,0,0.756656283737142,0.0343140896402799,-30.9725701802351;0.256127472105004,-78.4816875337030,0,0.424108014754793,0.254006097085642,-9.84553327948011;0,0,0,0,0,0.670320046035639];

Bd=[1.44391760017206e-06;-4.38901579232035e-09;4.38901579232036e-09;2.13261679566051e-05;6.97487101920980e-06;0];

Gd=[-0.205728181434440;0.00968483103669374;0.0308938523771156;-6.18883197642767;-1.40822087675432;0.329350383903707];

Dx=[0,0,1000,0,0,0;0,0,0,1,0,0;0,0,0,0,0,0];

Du=[-5;0;1];

Dw=0;

zmax=[-1e6 450 2.9e6];

zmin=[-1e8 350 2e6];

nx=6; nu=1; nw=1; nz=3; Sigw=0.02

Q=diag([1000 1000 1000 1000 1000 1000 ]);R=0.0005;

P=dare(Ad,Bd,Q,R); L=inv(R+Bd'*P*Bd)*Bd'*P*Ad; Qf=P;

S=[0;0;0;0;0;0];

[KSSV]

Yes works fine.

[Harleen Kaur]

Thank you so much.

[KSSV]

Replace:

L=inv(R+Bd'*P*Bd)*Bd'*P*Ad;

with:

L=(R+Bd'*P*Bd)\(Bd'*P*Ad);

[Harleen Kaur]

Will do that.