Changing size of variable inside Matlab function in SIMULINK

Question

0 个投票

Dear Matlab community

I am dealing with the following issue. I used the matlab function block inside SIMULINK which looks the following:

function du_llm2 = fcn(Yref,x,dztrack,PhiU2,PhiZ2,F2,H2,Qy2,Nc,dUmin,Umin,dUmax,Umax,Ymin,Ymax,M_du,M_u,C1)
    nu = 3;
    dU2 = zeros(nu*Nc,1);
    du_llm2 = zeros(nu,1);
    M_y  = [-PhiU2; PhiU2];  
    M = [M_u;M_du;M_y];
    gamma_du = [-dUmin; dUmax];
    gamma_u = [ -Umin + C1*du_llm2; Umax - C1*du_llm2];
    gamma_y = [ -Ymin + F2*x + PhiZ2*dztrack; Ymax - F2*x - PhiZ2*dztrack]; 
    gamma = [gamma_u;gamma_du;gamma_y];
    
    % % % % % % % % % % % % % % % % % % 
    M(isnan(gamma),:) = []; 
    gamma(isnan(gamma)) = [];
    % % % % % % % % % % % % % % % % % % 
    coder.extrinsic('quadprog'); 
    my_opt = optimoptions('quadprog');
%     my_opt.Algorithm = 'interior-point-convex';
    my_opt.Algorithm = 'active-set';
    my_opt.Display = 'off';
    % my_opt.MaxIterations = 1000;
    % my_opt.OptimalityTolerance = 1e-5;
    % my_opt.ConstraintTolerance = 1e-5;
    f2 = -PhiU2' * Qy2 * ( Yref - F2*x - PhiZ2*dztrack);
%     dU = -H2\f; % -inv(H)*f
    dU2 = quadprog(H2,f2,M,gamma,[],[],[],[],dU2,my_opt);
    du_llm2 = dU2(1:nu,1);
    
end
    

Basically, the problem here is that the size of variables inside the matlab function in SIMULINK are not allowed to change but cleaning M and gamma of the NaN values is essential for the further process. Is there any way to circumvent that?