The former information helps reduce the range for the scheduling variables. The latter
helps pick an optimal resolution (spacing) of samples in the scheduling space.Let us plot the actual trajectory of scheduling variables in the previous simulation against
the backdrop of gridded scheduling space. The outputs were logged via their
scopes (contained inside the Compare Responses block of scdairframeLPV).Stable = false(nA,nV);
for ct = 1:nA*nV
Stable(ct ) = isstable(G(:,:,ct));
end
alpha_trajectory = Alpha_V_Data.signals(1).values(:,1);
V_trajectory = Alpha_V_Data.signals(2).values(:,1);plot(alpha(Stable)*180/pi,V(Stable),'g.',...
alpha(~Stable)*180/pi,V(~Stable),'k.',...
alpha_trajectory,V_trajectory,'r.')
title('Trajectory of scheduling variables')
xlabel('\alpha'); ylabel('V')
legend('Stable locations','Unstable locations','Actual trajectory')3 Batch Linearization