|
|
Re: cognitive radio
Posted:
May 8, 2010 12:51 PM
|
|
"Vivi Sera" <vivi.seraphic@gmail.com> wrote in message <hrmljm$htd$1@fred.mathworks.com>...
> I am doing a term project in spectrum sensing as well, probably focus on cooperative schemes, but will start with the conventional one, such as energy detection, matched filter, etc etc. If any of you have some good scripts for me to start off, would it be okay for you to share with me, please? I really appreciate it!
hai iam sesham srinu from university of hyderabad,india
%Entropy estimation of a stochastic signal for primary user detection for the matched filter case
clear all;clc;
%**************************************Input parameters**********************************************************************************
b = [ 1 0 0 1 1]; % Input bit stream
N=[16 32 64];
for j=1:length(N)
snr_db=0:1:20;
g=length(snr_db);
%*************************************BPSK signal generation, Matched filtering and Sampling********************************************
for z=1:length(snr_db)
[signal]=bpsk(b); %BPSK signal generation
signal_fad=b_fad(signal); %fading
[s_matched signal_noise noise]=match(signal_fad,snr_db(z)); % matched filtered output of Qpsk signal
%************************************Entropy anlaysis after matched filter output is sampled********************************************
L=N(j); %No.of bins
s=s_matched(1:N(j)); % N no.of samples Sampled output of matched filter
x=[ ];
x=s(1:N(j)); %block of N samples of matched filter to make decision
noise=noise(1:N(j));
K=4;
sq=sum(x.^2)/N(j);
sigma = sqrt(sum(x.^2)/N(j));
%**********************************checking for chebyeshev inequality*******************************************************************
count=0;
for i=1:N(j)
if(abs((x(i)-mean(x)))>(K*sigma))
count=count+1;
else
count=count;
end
end
if ((count/N(j))<= (1/(K^2)))
%display('Design value of K is correct ');
else
%display('Design value of K is incorrect and try for another value');
end
% %***********************************Histogram of x**************************************************************************************
v1=(2*K*sigma)/(L); % Width of each bin
v=-K*sigma:v1 :K*sigma; % a vector specifying the centres of bins from where to start and stop.
%***********************************Checking for no.of values of x fallong inside the kth bin********************************************
pa11=0;
pa01=0; y1=0;e(z)=0;
for k=1:L;
number_samples=0;
lk(k)=v(k)-(v1/2);
lkplus1(k)=v(k)+(v1/2);
for i=1:N(j)
if((lk(k)<x(i))& (x(i)<lkplus1(k)))
number_samples=number_samples+1;
else
number_samples=number_samples+eps;
end
end
nk=number_samples;
y1=y1+number_samples;
%----------------------------------------Entropy caculation------------
if(nk~=0)
nk=nk;
else
nk=nk+eps;
end
p(k) = -((nk/N(j))*log2(nk/N(j)));
e(z)= e(z)+p(k) ; %Entropy estimate
%********************************************computing pk(a)********************************************************************************
a1=snr_db(z);
a0=0;
pk1=0;pk0=0; M=2; u(1)=1;u(2)=-1;
for m=1:M
h1(m)=0;h2(m)=0;h3(m)=0;h4(m)=0;
h1(m)=(lk(k)-u(m)*sqrt(a1/(1+a1)))*(1+a1);
h2(m)=(lkplus1(k)-u(m)*sqrt(a1/(1+a1)))*(1+a1);
h3(m)=(lk(k)-u(m)*sqrt(a0/(1+a0)))*(1+a0);
h4(m)= (lkplus1(k)-u(m)*sqrt(a0/(1+a0)))*(1+a0);
pk1=pk1+[[qfunc(h1(m))]-[qfunc(h2(m))]];
pk0=pk0+[[qfunc(h3(m))]-[qfunc(h4(m))]];
end
pka1= (1/M)*pk1+eps;
pka0= (1/M)*pk0+eps;
%*********************************************computing log-likelihood ratio***************************************************************
pa11=pa11+(nk/N(j))*[log2(pka1)-log2(pka0)];
end
le(z)=pa11;%Log-likelihood ratio of entropy
end
e1((j+(j-1)*20):j*21)=e;
le1((j+(j-1)*20):j*21)=le;
end
plot(snr_db,e1(1:21),'-b*',snr_db,e1(22:42),'-ro',snr_db,e1(43:63),'-g.');
xlabel('SNR in db');
ylabel('Entropy');
title('Normal entropy');
h2=legend('N=16 BPSK','N=32 BPSK','N=64 BPSK',1);
figure
plot(snr_db,le1(1:21),'-b*',snr_db,le1(22:42),'-ro',snr_db,le1(43:63),'-g.');
xlabel('SNR in db');
ylabel('Log-likelihood Entropy');
title('log-likelihood entropy')
h2=legend('N=16 BPSK','N=32 BPSK','N=64 BPSK',1);
|
|
