Error Control Code - Linear Block Codes
Aim:
To study the BER performance of Linear Block Code
Requirements:
MATLAB software,
Desktop Computer
Formula:
Bit error rate = No. of bits in error after detection / No. of bits transmitted
Theory:
Block Diagram:
Algorithm:
1. Generate binary random bits to the desired length
2. Convert them into block of 7 bits to form it as message block
3. Generate generator matrix
4. Multiply the message block with generator matrix and get codeword blocks
5. Convert back codeword blocks into bit stream.
6. Generate noise to yield the required SNR with respect to the energy in the codeword bit stream.
7. Decode the codeword without error detection and correction. Compare the decoded bits with
transmitted and calculate BER without error control
8. Convert the codeword bit stream into blocks again. Apply parity check matrix to calculate
syndrome
9. Identify the respective error pattern and add with codeword to generate error corrected codeword
10. Separate message words and get the output message bit stream. Compare this input message bit
stream and calculate BER with error control
11. Repeat the steps 6-10 for different SNR and plot BER Vs SNR curve with and without error
control coding
Program:
clc; clear all; close all;
n=7; k=4;
blk_nos=2;
Nbits=blk_nos*k; % Number of bits in simulation
snrdB=[0:10];
% P= [ 1 1 0 ; 0 1 1 ; 1 1 1 ; 1 0 1]; % Parity Matrix
P=[ 1 1 1; 1 1 0; 1 0 1; 0 1 1];
In= eye(4); % Identity Matrix
Ink=eye(3);
G=[In,P]; % Generator Matrix
H=[P',Ink ]; % Parity Check Matrix
bs=randint(1, Nbits, 2); % message bit generation
m=reshape(bs, k,[] )'; % Blocking of bits into 4 bits in each block
x=mod( m*G, 2); % Generating Codeword
ct=reshape(x', 1,[]); % Transmitted bit sequence
ctx=ct;
ctx(ct==0)=-1;
%% For error calculation
errpat=[ 0 0 0 0 0 0 0 ;
1000000;
0100000;
0010000;
0001000;
0000100;
0000010;
0 0 0 0 0 0 1 ];
syndrom=errpat*H';
%% Transmitting bits thr AWGN channel
for ii=1:length(snrdB)
snr(ii)=10^(snrdB(ii)/10)
r=awgn(ctx, snr(ii), 'measured');
cr=r;
cr(r>=0)=1;
cr(r<=0)=0;
y=reshape(cr', n, [])';
s=mod( y*H', 2);
e0=[];
for jj=1:blk_nos
for kk=1:2^(n-k)
if(syndrom(kk,:)==s(jj,:))
ro_match=kk;
break;
end
end
e0=[e0; errpat(ro_match,:)];
end
xcat=mod( y+e0, 2); % 1 bit error correction
BER_wo_EC(ii)=length(find(ct~=cr))/length(ct)
BER_with_EC(ii)=length(find(x~=xcat))/length(ct)
end
semilogy(snr, BER_with_EC, 'r*-', snr, BER_wo_EC, 'b-+')
legend('BER_with_EC', 'BER_wo_EC,');
xlabel('SNR (dB)'); ylabel('BER'); title('BER Performance of Linear Block Code');
Observation:
Number of blocks on simulation : 20000;
Number of bits used: 140000
Message word length = 4
Codeword length = 7
Generator Matrix=
Parity Check Matrix =
SNR:
BER (w/o EC):
BER (with EC):
