Adding AWGN to the OFDM PLC generator

Hi,

My code below is about OFDM PLC generator using bpsk. I am stuck with adding the AWGN to the OFDM PLC code below, I appreciate any help regarding the code (where can I exactly add the code and what is the code that can be added to generate the AWGN and turbo codes). Appreciate your kind help on this regard.

      %generation of data
  l = 16384*32;   % 1Mbps this is 1 mega bit of data from here we determine the bit rate of the system
  lpf = l/32;        % here we give the no. of user and data length of each user ..... u can use 64 as well as 128 that means 2 ki power(no. of user)
  data = randint(l,1,2);    % it generate randint(no. of row, no. of column, (0 and 1)) this is for bpsk if u want for QAM -16 then write 16 
  % data get generated

% serial to parallel
p_data = reshape(data,32,[]); %here reshape(matrix, no. of row, no. of column)
% serial to parallel ends

%modulation scheme 
%here i use bpsk u can use qam as qammod(matrix, no. of level same as that in line no. 9 )
m_data = pskmod(p_data,2,pi); % binary phase shift key it best and mostly use that why i have used which give 0 as -1 and 1 as +1
% modulation ends

%now ifft
x_ifft = ifft(m_data);
%ifft ends

%cyclic prefix
p_cyc = [x_ifft(end-4:end,:) ; x_ifft]; % here i concatenate the last L-1 bit with whole matrix so that it act as a cyclic prefix
%its use is just to avoid inter symbol interfernce 
%her L is 6 (that is length of channel delay(6) if it is more u use just L-1 bit of cyclic prefix)
%ends of cyclic prefix

%now again parallel to serial bcoz we have to transmit and we cant transmit
%serially 
s_cyc = reshape(p_cyc,1,[]); % serially data
%end parallel to serial

%now every thing gets over now our aim is to transmit their now we design
%the channel which include multi path gain and channel noise and to
%transmit we have to do the linear convolution

%lets take channel to be random path gain
channel = randn(6,1);
hf = fft(channel,32); % just to get the impulse response of gain coefficent for each user if user is 32 then length is 32 if 64 thn use 64 length
%channel coefficent ends

%linear convolution
chnout = filter(channel,1,s_cyc);
%linear convolution ends

%generation of noise as here we are using asynchronous impulse noise for
%the case of power line communication
noiseq = randn(length(chnout),1);
tin = 1;
IAT = 10276;
tnoise = 34;
for ik = 1:51,
    for ii = tin:tin+IAT-1,
        impulse(ii) = imp((ii-tin)/tnoise);
        impamp(ii) = impulse(ii)*noiseq(ii)*10;
    end
    tin = tin+IAT;
end
clear noiseq
leng =length(impamp);
noiseq = [impamp zeros(1,length(chnout)-leng)].';

%noise generation

%In the case for the wireless channel

AWGN;

%noise generation ends

%receiver along with noise addition in channel as the stength of noise
%depend on the signal to noise ratio that we are going to take
sereq  = []; 
power = mean(mean(abs(p_cyc).^2));

for il = 1:30 ,  % signal to noise ratio ranges from 1 to 30
    eb2n(il) = il;
    eb2n_num(il) = 10^(eb2n(il)/10);
    var(il) = power/(2*eb2n_num(il)); %variance of the noise
    signois(il) = sqrt(var(il)); % strength of noise
    impnois = signois(il).*noiseq;
    yout = chnout+impnois.';  % addition of noise in transmited signal

      %serial to parallel conversion%
      pout = reshape(yout,37,[]);
      %serial to parallel over%

      %removal of cyclic prefix
      pncyc = pout(6:end,:);
      %removal of cyclic prefix ends

      %fft%
      x_fft = fft(pncyc);
      %fft over%

      %just to remove the effect of channel coefficent
      z_data = inv(diag(hf))*x_fft;
      %ends

      %decision making%
      %parallel to serial%
      h_data = reshape(z_data,1,[]).';
      %serial to parallel ends

      %demodulation
      dec = pskdemod(h_data,2,pi);
      %demodulation ends

      %comparison nd to determine no. of bit in error
      sereq = [sereq sum(dec~=data)/lpf];
      %ends
  end

%plot of bit error rate as compare with signal
semilogy(eb2n,sereq,'k-*');        %semilogy function is used as semilogy(xaxis , yaxis,'colour and shape of graph') and semilogy will make yaxis as logarithmic
%ends

Answer 1

[Ayaz Hussain]

impulse(ii) = imp((ii-tin)/tnoise); instead of impulse(ii) = double((ii-tin)/tnoise);