Eigenvalues performance for uncorrelated signals

Hi everyone,

I want to study the eigenvalues performance for two source signals, one of the signals have theta=0, while the other one have theta that change from -45 to 45. I implemented the following code

N_sig=2;

Theta1=[0]/180*pi; %%% first Source angle

Theta2=[-45:1:45]/180*pi; %% second Source angle

doa=[Theta1 Theta2]/180*pi; %Direction of arrival

P=[0:9]; %% number of antennas

lambda = 1; %% wavelength

d=lambda/2;%Element spacing

snr=10;%SNR

sigma2 = 0.01; % Noise variance

s_var= [10]; %% signal variance

center_freq=10^6;

Snap=100; %% number of snapshots

% % generating noise

noise=(1/sqrt(2))*(randn(length(P),Snap)+j*randn(length(P),Snap));

noise_std=10^(-snr/20);

n=noise_std*noise;

% generated signal

F=randn(N_sig,Snap)*pi;

Signal=sqrt(s_var).*exp(j*2*F);

% generate Data matrix

A=zeros(N_sig,length(P)); %To create a matrix with N_sign row and P column

for kk=1:N_sig

A(kk,:)=exp(j*2*pi*d*P'*sin(doa(kk))/lambda); % steering matrix

end

A=A';

x=A*Signal+n; % received signal

% % Data covariance matrix

R=x*x';

[V, D] = eig(R);

eigs_sorted = sort(diag(D),'descend');

DD=10*log10(eigs_sorted);

I have tried several ways, but I don't get the same result as in the figure shown below.

Any help I realy apperciate it.

[Yazan]

You forgot to define doa in your code.

[fatoma Saml]

I define it in the code, but it was in captial letter, now I edit it.