rls.m

智能天线自适应波束形成算法的研究,进行了基本算法的改进与运算,大家可以进行参考,以便更好的学习.

clear,clc
m=8;                                     % sensors
n=2;                                     % sources
theta=[-20 0];                           % in angle
d=1/2;                                   % 1/2 lambada
N=200;                                   % samples
L=100;                                   % resolution in [-90' 90']
Meann=0;                                 % mean of noise
varn=1;                                  % variance of noise
SNR=10; 
INR=10; 
rvar1=sqrt(varn) * 10^(SNR/20);          % variance of signal
rvar2=sqrt(varn) * 10^(INR/20);          % variance of interference

% generate the source signals
s=[rvar1*exp(j*2*pi*50*0.001*[0:N-1])
   rvar2*exp(j*2*pi*(100*0.001*[0:N-1]+rand))];
% generate the A matrix
A=exp(-j*2*pi*d*[0:m-1].'*sin(theta*pi/180));
% generate the noise component
e=sqrt(varn/2)*(randn(m,N)+j*randn(m,N));
% generate the ULA data
Y=A*s+e;

% initialize weight matrix and associated parameters for RLS predictor
de=s(1,:);
w = zeros(m, 1);
lambda=0.75;
delta=1e-2;
P=1/delta*eye(m);
for k = 1:N
    v=P*Y(:,k);
    u=1/lambda*v/(1+1/lambda*Y(:,k)'*v);
    e(k)=de(k)-w'*Y(:,k);
    w=w+u*conj(e(k));
    P=1/lambda*(eye(m)-u*Y(:,k)')*P;
end 

% beamforming using the RLS method
beam=zeros(1,L);
for i = 1 : L
   a=exp(-j*2*pi*d*[0:m-1].'*sin(-pi/2 + pi*(i-1)/L));
   beam(i)=20*log10(abs(w'*a));
end

% plotting command followed
figure
angle=-90:180/L:(90-180/L);
plot(angle,beam);
xlabel('angle');
ylabel('幅度响应/dB');
figure
for k = 1:N
    en(k)=(abs(e(k))).^2;
end
semilogy(en);
xlabel('n');
ylabel('e^{2}(n)');