Important Terminology :
Let us choose an integer n m and one-to-one function
e :Bm Bn .
Encoding Function :
The function e is called an (m, n) encoding function. It means that every word in Bm as a word in Bn.
Code word :
If bBm then e(b) is called the code word
Weight :
For x Bn the number of 1’s in x is called the weight of x and is
denoted by x .
e.g. i) x 10011B5 w x 3
ii) x 001B3 w x
x y Let x, yBn , then x y is a sequence of length n that
has 1’s in those positions x & y differ and has O’s in those positions x & y are the same. i.e. The operation + is defined as 0 + 0 = 0 0 + 1
= 1 1 + 1
= 0 1 + 0 = 1
e.g. if x, yB5
x 00101, y 10110
x y 10011
w (x y) 3
Hamming Distance :
Let x, yBm. The Hamming Distance x, y between x and y is the weight of x y . It is denoted by x y . e.g. Hamming distance between x & y can be calculated as follows : if x = 110110, y = 000101 x y = 110011 so x y = 4.
Minimum distance :
Let x, yBn.then minimum distance = min d x, y / x, yBn.
Let x1, x2 xn are the code words, let any xi, i 1 n is a transmitted word and y be the corresponding received word. Then y xk if dxk, yis the minimum distane for k = 1, 2, --- n. This criteria is known as minimum distance criteria.
Detection of errors :
Let e : Bm Bn m n is an encoding function then if minimum distane of e is ( k + 1) then it can detect k or fewer errors.
Correction of errors :
Let e : Bm Bn m n is an encoding function then if minimum distance of e is (2k + 1) then it can correct k or fewer errors.
Weight of a code word : It is the number of 1’s present in the given code word.
Hamming distance between two code words : Let x x1 x2 ... xm and y y1 y2 ... ym be two code words. The Hamming distance between them, x, y , is the number of occurrences such that xi yi for i 1, m .
Example 1 : Define Hamming distance. Find the Hamming distance between the codes.
(a) x 010000, y 000101 (b) x 001100, y 010110
Solution : Hamming distance :
(a) x, y x y 010000 000101 010101 3
(b) x, y
x y
001100 010110
011010 3
Example 2 : Let d be the 4, 3 decoding function defined by
d : B4 B3 . If y y1 y2 ... ym1 , d y y1 y2 ... ym . Determine d y for the word y is B4 .
(a) y 0110 (b) y 1011
Solution : (a) d y 011 (b) d y 101
Example 3 : Let d : B6 B2 be a decoding function defined by for
y y1 y2 ... y6 . Then d y z1 z2 .
where
zi 1 if y1, yi2 , yi4 has at least two 1’s.
0 if y1, yi2 , yi4 has less than two 1’s. Determine d y for the word y in B6 .
(a) y 111011 (b) y 010100
Solution : (a) d y 11 (b) d y 01
Example 4 : The following encoding function f : Bm Bm1 is called the parity m, m 1 check code. If b b1 b2 ... bm Bm , define
e b b1 b2 ... bm bm1
where
bm1 0 if
= 1 if
b is even. b is odd.
Find eb if (a) b 01010 (b) b 01110
Solution : (a) eb 010100 (b) eb 011101
Example 5 : Let e : B2 B6 is an (2,6) encoding function defined as e(00) = 000000, e(01) = 011101
e(10) = 001110, e(11) = 111111
Find minimum distance.
How many errors can e detect?
How many errors can e correts?
Solution : Let x0, x1, x2, x3 B6 where x0 000000, x1 011101, x2 001110, x3 111111
w x0 x1 w 011101 4 w x0 x2 w 001110 3 w x0 x3 w 111111 6 w x1 x2 w 010011 3 w x1 x3 w 100010 w x2 x3 w 110001 3 Minimum distance = e = 2
Minimum distance = 2
An encoding function e can detect k or fewer errors if the minimum distance is k + 1. k 1 2k 1
The function can detect 1 or fewer (i.e. 0) error.
e can correct k or fewer error if minimum distance is 2k + 1.
2k + 1 = 2
k = 1
2
GROUP CODE :
An m, n encoding function e : Bm Bn is called a group code if range of e is a subgroup of Bn. i.e. (Ran (e), ) is a group.
Since Ran (e) CB n and if (Ran (e), ) is a group then Ran(e) is a
subgroup of Bn. If an encoding function e : Bm Bn (n < n) is a group code, then the minimum distance of e is the minimum weight of a nonzero codeword.
.
DECODING AND ERROR CORRECTION :
Consider an m, n encoding function e : Bm Bn , we require an (n,m) decoding function associate with e as d : Bn Bm .
The method to determine a decoding function d is called maximum likelihood technique.
Since Bm 2m .
Let xk Bm be a codeword, k = 1, 2, ---m and the received word is y then. Min 1 k 2m d xk, y d xi, y for same i then xi is a codeword which is closest to y. If minimum distance is not unique then select on priority
MAXIMUM LIKELIHOOD TECHNIQUE :
Given an m, n encoding function e : Bm Bn , we often need to determine an n, m decoding function d : Bn Bm associated with e. We now discuss a method, called the maximum likelihood techniques, for determining a decoding function d for a given e. Since Bm has 2m
elements, there are 2 m
a fixed order.
code words in Bn . We first list the code words in
1 2
2m
x , x , ..., x
min
1 i2 m
xi , x1 xs , x1
That is, xs is a code word that is closest to x1 , and the first in the list. If xs e b , we define the maximum likelihood decoding function d associated with e by
d xt b
Observe that d depends on the particular order in which the code words in e Bn are listed. If the code words are listed in a different
order, we may obtain, a different likelihood decoding function d associated with e.
Theorem : Suppose that e is an m, n encoding function and d is a maximum likelihood decoding function associated with e. Then e, d can correct k or fewer errors if and only if the minimum distance of e is at least 2k 1 .
1 1 0
0 1 1
Do'stlaringiz bilan baham: |