doesnt make any sense plot (RLS) !!

original system:

y(k)=2.7*y(k-1)-1.8*y(k-2)+u(k-1)+5*u(k-2)+6*u(k-3);

it's required to get the parameters using recursive extended least square..so my code is:

%Recursive Extended Least square
u=ScopeData2(:,2);
 P=1000*eye(4);
 theta=[0 0 0 0]';
y=zeros(500,1);
for k=4:500
y(k)=2.7*y(k-1)-1.8*y(k-2)+u(k-1)+5*u(k-2)+6*u(k-3);
 phai=[y(k-1) -y(k-2) u(k-2) u(k-3)]';
 P=P-((P*phai*phai')*inv(1+(phai'*P*phai)));
 e(k)=y(k,1)-phai'*theta(k-1);
  theta=theta+P*phai*e(k);
 a1(k)=theta(1);
 a2(k)=theta(2);
 b1(k)=theta(3);
 b2(k)=theta(4);

end

when i run program and plot values of a1,a2,b1,b2..it gives me NAN instead of converges to their true values :S

any help would be greaaaat .

A

[Marwa Ali]

u is input signal ...i used square wave!