0% found this document useful (0 votes)
223 views9 pages

Group Codes Continue

The document discusses error correcting codes and properties of group codes. It contains the following key points: 1) A group code is an encoding function e: Bm -> Bn where the range of e forms a subgroup of the group <Bn, e>. 2) For a group code, the minimum distance is equal to the minimum weight of a non-zero codeword. 3) Operations on Boolean matrices like modulo-2 sum and modulo-2 product are defined, and properties like (D * E) * F = (D * F) * (E * F) are proven. 4) It is shown that a vector x satisfies the parity check equation x*H

Uploaded by

Thakre Shubham
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
223 views9 pages

Group Codes Continue

The document discusses error correcting codes and properties of group codes. It contains the following key points: 1) A group code is an encoding function e: Bm -> Bn where the range of e forms a subgroup of the group <Bn, e>. 2) For a group code, the minimum distance is equal to the minimum weight of a non-zero codeword. 3) Operations on Boolean matrices like modulo-2 sum and modulo-2 product are defined, and properties like (D * E) * F = (D * F) * (E * F) are proven. 4) It is shown that a vector x satisfies the parity check equation x*H

Uploaded by

Thakre Shubham
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 9

Exam ple 4 :

a) Find the mini mum dista nce of the (2, 4) enco ding func
tion e, give n belo w :
e(OO) = 0000~ e(10) = 0110, e(0l} = 1011, e(ll) = 1100.
b) How many error s will e detec t?
Solution : Let x = e(OO) = 0000, x = e(l0) = _0110
1 2
x3 = e(0l) . = 1011, x_. = e(ll) = 1100
The dista nce betw een x and y is I x e y I
Here, I "i 6l x2 I = I 0000 EB 0110 I = I 0110 I = 2,
I "i 6l ~ I = I 0000 EB 1011 I = I 1011 I =. 3,
I "i @ x_. I = I 0000 EB 1100 I = 11100 I, = 2,
I x2 @ x 3 I = I 0110 EB 1011 I = 11101 I = 3,
I x2 6l x_. I = I 0110 EB 1100 I = 11010 I = 2,
I~ 6l x_. I = I 1011 EB 1100 I = I 0111 I = 3
,a) From abov e we conc lude that the mini mum dista nce
of e is 2.
b) Since e can detec t k or fewe r error iff its mini mum dista
nce is at least k + 1.
We have 2 ~ k + 1 or l< ~ 1
:. e can detec t 1 or fewe r error .
:. e can detec t k = 1 or fewe r error .

Example 5 : Fi,nd the dista nce betw een. x and y


a) x = 00111001 ~d y = 10101001.
b) x = 11010010 and y = 00100111
c) x = 1100010 and y = 1010001
d) x = 0100110 and y == 1010001
Solu tion : Dista nce betw een x and y is I x e y I
a) I x e y I = I 00111001 e 10101001 I = I 10010000 I =
i
b) I x e y I = I 11010010 e 00100111 I = I 11110101 I
= 6
c) I x e y I = I 1100010 e 1010001 I = I 0110011 I = 4
d) I x e y I = I 0100110 e 0110010 I = I 0010100 I = 2

Exam ple 6: Cons ider (2, 6) enco ding funct io_n e as e(00)
= 000000, e(l0) =
e(0l} = 011110, e(ll) = 111000 1010 10
. - '
a) Find the minimum dista nce.
b} How ~ y error s will e detec t?
Solution : Cons ider the (2, 6) enco ding func tion e as : "i = e(00)
= 000000, x =
e(10) = 101010, x3 = e(l0) = 011110, x_. = e(ll) = 111000 2

I x1 e x2 I = I 000000 e 101010 I = I 101010 I = 3


I X, e x~ I = l 000000 e 011 110 I :::: I Ol 1110 I = 4
I x, e x, t = I 000000 a 111000 J = I 111000 t • 3
I ~ e xJ I = I 101010 s 011110 t = I 110100 I : 3
t Xa e x ~ I = f 101010 ~ 111000 I = I 01001 O I = 2
I ~ e x" I = I 011110 s 111<XXJ I = I 100110 I = 3
a) The min imu m distance is 2
distance is at leas t k + 1.
b) As e can detect k or fewer erro rs iff its min imu m
\Ve have 2 ~ k + 1 or k ~ 1
:. e can dete ct 1 or fewe r erro r.
e as e(OOJ) = CXKXXXXO.l, e(OOl)
Exa mpl e 7 : Con side r the (3, 9) enco ding func tion
010011010,
V = 011100101, e(010) = 01010100), e{Ol l} = 11001cxxn, e(100) =
e(lO l )= 111101011, e(110) = 001011CXXJ, e(111) = 11cn xn1 1.
a) Find min imu m distance.
b) How many erro rs will e detect?
n by
Solu tion : Let e _be the (3,9) enco ding function give
"i = e (OOJ) = OOCXXXXXX)
~ = e{001) =
011100101, ~ = e(Ol0) = 010101(0)
Xi = e(011) = 11001CXXJ1, x..> = e(100) = 01001101

0

X. = e{101) = 111101011, x, = e(l 10) = 0010 11(0 ) -

x, = e(111} = 11CXXXJ111 ~
The distance betw een x and y is I x ~ y I
I ~ a "2 I = 5, I ~ s x 3 I = 3, I ~ s x: I = 4,
I ~ s x l = 4, f x_ e xr. I = 7, , ~ e x.. J = 3,
3
I Xi e x_ I = 5, f ~ a x 3 I = 4, I x1 s x~ I = 5,
I ~ a x.: f = 6, t ~ e x, 1= 3, 1 ~ & x, ? = 6,
f x, e Xa I = •t I x 3 e x, I = 5. l ~ ~ X; I = 3,
I x} 8 x,. I - 4, I xl s x_ I = 4, I x 3 e X. I = 6,
f x, a x3 f = 4, I x. a x. I = 5,. i x. e x, t = 5,
~ _ f .: 4., I x . a .l• t = 5, t .)._,.J a 'L l = 3.
l x ' 9 -... I
.)

f ~ 8 .~ f ~ 5, l x, ii ~ l = b , i "• e ~ ~ = -1,
f x,0 x. l ; '8
a) from above the rrunimum distance ol ~ is 3
b) The enco ding function e can defect t or fewer errors iff its min imu m dista nce
is adeast k ,.. 1.
\\fe have 3 ~ k + 1 or k ~ l
·. e can dett..~ t 2 or ·fe'\1a•t T e,rro rs
Group codes 4.4: An (m, n) encoding function e: Bm ➔ B" is called as a group codr
if e (Bm) = {e (b) I b E Bm} = Ran { e} is a subgrou p of group <B" , e >
Note : H~e ~e ~ow !:}lat N is subgroup of group B" if a) the identity of 8" is in N
b) if x and y belong to N then x e y E N c) if x is in N the its inverse is in N.
Theorem : Let e : Bm ➔ Bn be a group code. Then the minimum distance of e is the
minimum ~eight of a non zero code. , • -. 1 ~ o,f fl M 7. ~ r,,.
Proof : Let e : Bm ➔ Bn be a group code d,, 1 ,; ~ n 6" f""" \ f'\ · v,.)-t} ~ ,,, ·( , d Q 'rJ t/'<.
Let 6 be the minimum distance of the group code. Assume that o = b(x, y)
C~\ c,1. ~ w h ere x and y are distinct code · h t Of ~
words. Also let n be the minimum we1g
r ';.. \ 1 ~
", a no~ _zero code word and suppose tliat n = I z I for a code word z. ~ C!-,,,...
u .: e
is a group code, x e y is a .non zero code word.
Thus 6 = 6(x, y) = I x e y I ~ n ....... (1)
on the otherhan d, since o and z are distinct code words.
'

n = I z I = I z e o I = o (z, o ).~ o .. :... (2)


from (1) & (2) o ;!: n & n ;!: o
- n =o
Hence the minimum distance of e is the minimum weight of a non zero code
word.
Definition : (1) _consider _the set B with operation + defined in .table as
+ . or 1\. 1

0 0 1
1 1 0

Now let D = .[d-]



and E = [elJ..] be mxn Boolean matrices. We define mod 2 sum
of D and E as the Boolean matrix De E which is also mxn matrix F = [F..] where
lJ
f, = d• + e., 1 ~ i ~ m and 1 ~ j ~ ·n Here note that if F = D e E then f ii = o
when both dii and eii are zero or both are 1. ·
(2) Conside r the set B with binary operation given in the table as

* 0 1
0 0 0
1 0 1

Here B is the unique ·Boolean algebra with ·two elements ,.


Let D == (dij] be mxp Boolean ma~ & E = [elk] be a ·p xn Boolean matrix. We
define the mod 2 B~lean product D e E = F = [fnJ mxn where
l
~.
f,, : f d• .eik : d" .e,.
J= l
+du. e,. + ....... + d,,.e"' , 1 :,;; ,;; m, 1 ,; i :s; m, i ,; j ,; n.

Boolean matrix.
Theo rem : Let D and E be mxp Boolean matrices and let F be a p:xn
0

Then (D e E) * F = (D * F) e (E * F)°
Theo rem : Let x = YiYi-···Y m~····xr ... Xr E B". Then x * H = if and o
only if x = eH(b)
for some b E Bm. \..J \.,_ •U .. " ~ ·, _, ya--r a(
~,.o c"l-.. , N' &) ~ V o! 0\- ,,d.. Ut

.
1" vY

Proof : Let X = Y1Y2····Ym~Xi-···xr E 8".


Let H be an nxr Boolean matrix as given below :
h 11 h 12 h 1r Here last r rows form the r x r ident ity
h 21 h ll hb ma trix and H is called as a parity
check matrix.
h ml hm2 hmr
1 0 0
H= 0 1 0

0 0 1

First assum e that x * H = 0 · This giv~s


y,l\, +y,h,, +. . +.~mh°';: ~ ~ 0) .... (1)
Y1J\2 + Yfln Y
+ ... . m m2 2 -

: h+ +yhmr +x= O
Y11\r + Yr"a .... m • r
. th f ' t equat ion in (1) IS as ..... (2)
We wnte e rrs u..:t
....i--e a = y r"n + Yfln + .... + Ymhmr
, + x1 = 0 , .Yn1c1. h

' · a~ a + (a + x1 ) = a + 0 (': ad~ g a to both sides of (2))


- a
~ (a + a ) + x1 -
(': a + a =0

~ Xi = a. d e for each row, therefore


. -
This can be on
j, + + y h . for 1 - 1,2 .... r
- h + Yr"2i ....
b
Ill ma
_ b .
x. - Yr.. --ii
I
. = b1, yz -- z, ....., y,,. - m'
Talon g Y1 atisfy the equat ion as
we see that x,, X2 .•. Xm 5 . -

-bh . h . for 1 - 1,2 .... r )


+ b zh 2i. + .... + b "'nu
X - .. , . b h
i 1 I • b h + bfln + .... + m m1
- ..... (3)
whic h g:iv~ Xi - 1 .. --i1 b h
= b 1l\i + bPzz + .... + m °'2
Xi = b h + bPzr + ···· +bh m mr
xn 1'.. '1,
• I
-.r
11

1 v
H~n11 h • 1l, 1\ , J'"' 1l, l1\ <: ·1~ 11'

H~n., ~u : »'" ➔ B" i~ ,lrfinClcl by


~ " • --~II (b) • h,b2hy ... b,nx,x, .... x, for b = b,b, ....h"' E IJI"

wh,""' xi, ~.... 'l<, ~,re as in (3) .


~-\-'1''-'_\rsclv ~ppost? tha t an encoding function
~ : 8 111 - B" is defined by x = e 11 (b) for some b = b 1b2•••• b,. E 8 \ whel'e X1X,····x,
11

at-e as in (3). Here equations in (3) can be rewritten by adding x 1 to both. sides
of th e 1·th equation,.
1. -- }. , 2 .•.. r as

b1h.1 + h 2~1 + ···· + bmhmt + X, = Xl + x, = O


blh,2 + h 2h22 + .... + bmh,n2 + x2 = x, + x2 = 0

h r + .... + b mh m, + xr = xr + xr =O
b 1h• '1 r + b 2"''1
which is m.erely as x * H = o
Ex.ample 8 : Show that the (2, 5) encoding {unction e : Bl - 8 ~ defint.:'Ci by
e(OO) • 00000, e(0l) a: 01110
e(l0) • 10101, e(11) • 11011, is a group code.
Solution : Consider the set of all codewords as e (B•) •
N • (00000, 01110, 10101 , 1101.1) • ( o, a, h, c)
where o•
00000, a • 01110, b • 10101 , , • 11 011 .
:. Compnsition tab}(, for <N, <D > is

$ () a b L

5 0 0 b ('

a a 0 (' h
h b ('
0 •
h a 0 .\ '
C C
H..>r,·, I tjJ j tt (n)i.)ll -
- t)
b ,111 h .. l) 4 t • ,· - ll
aG..h ... (
1 ~ < • 10101 Q h
b • t .-. OJ 110 ~

ckarly O ... (XX(X) it ,Je1,Uf)• ~ N 41\J t'\.., ••) · d1.•nlt>-nt of N l• ,~ '-'Wn in\'~fW'
Alsa aU d~ ,-n~ "' ,h..- •~t»..- liru tlw ,.. l~u~1, h .:,I N ll~~ , N# ~ > 1$ a
subgroup of 8 , ,u\11 1lils> ttw, th,.. in i••n ,·t\\.v1.hn>~ hu"- u.ou ,. L'> a ~,uup o.lJl
Example 9 : Show _that the (3, 7) encoding functi~n
7
➔ B defined by e(OOO) = 0000000
3
e : B
e(OOl) = 0010110, e(OlO) = 0101000, e(Oll) = 0111110
e(lOO) = 1000101, e(101) = 1010011, e(llO) = 1101_101
e(lll) = 1111011, ~ a group code.
Solution : Consider the set of code words as e(B7) := N = {0000000, 0010110,
0101000, 1000101, 0111110, 1010011, 1101101, 1111011 }
= {0, a, b, c, d, e, f, g }, where
0 = 0000000, a = 0010110, b = 0101000, c = 1000101, d = 0111110, e =
1010011, f = 110110~, g = 1111011
composition table for <N, e > is as below :
e 0 a . b C d e f g
0 0 a b C d e f g
a a 0 -d e b C g f
b b d 0 f a g C e
C C e f 0 g a b d
d d b a g ·O f e C

e e C g a f 0 d ·b
f g C b e d 0 a
f
f e d C b a 0
g g
. the 1.denti·ty for N = e(B7) and every element of N 1s its own inverse
clearly O is
entries are ·the elements of N. Hence N = e(B7) is a
and from -table all the
7 that e is th~ group code.
subgroup Of B ' so - ·
Example 10 : · Compute
- a)
[ ~ I 0] [I : :] I 1 @ · 0

l l I
1 0 l
I 1 0 I O I
@ O I 0
O O I
0 I I O O I

Solution :
I t ] = [ I + I I + I O+ I]
I 11 w [ 01
Ol]in I I O+O l+l l+l
a) [O
- [o
- 0 0
0
~]
0 1 1
0 1 0
EB
b) 0 0 0 1 0
0 0 0

1+ 1 0+l 1+ 1 0 0
l+l l+0 0+l 0 1 I
=
= 0+0 0+l l+0 0 1
0+0 l+0 l+l 0 l 0

Example 11 : Compute


a) r 1 · ~]
.0 I
1
· I O • [~- :

b)

Solution ·: -
u~ :J •[: i i]
l] [lxO+lxl lxl+lxl lxl+lx0]
a) O
= Ix0+0xl lxl+0xl lxl+0x0
0x0+lxl . 0xl+lxl 0xl+lx0

= [i:~ : ~ : ~]
0+l 0+I 0+0
= [~
I
0 0~]
I

-b)
u :J ·[: I
0
0
0

[lxO+lxl+Oxl lxl+lx0+0x0 lxl+lxl+OxO]


l]
0x0+lxl+lxl 0xl+lx0+lx0 0xl+lxl+lx0
==
Jx0+0xl+lxl lxl+0xl+lx0 lxl+0xl+lx0
-
o+ ' + o ' + o+ o
[ O+ t + I 0 + 0 + 0
l+O+O
I+ I+ o
O+l+O
l+O+O
l
0 + ( 1- I

- [~ ~ :]
Example 12 :
0 I I
0 I l
0 0 be
Let H = 0 l 0
0 0 1
2 ➔ B5
a parit y check matri x Determine the (2, 5) group code funct ion eH : B
0 l l
0 I l
l O 0
Solution : Let H = be a parity check matrix and
0 1 0
0 0

eH : B2 ➔ B5 be an ~need ing function : Here


B2 ...:. { 00, 01, 10, 11}
Let b = b~ bi E B2, then eH(b) = b1b2 ~X2'S E B5

Wher e xl = bl .0 + b2.o =0
X2 = bl. 1 + b2. 1 = bl. + b2

X3 = + b-,. l -- b1 +b 2
- b .1
1 -
• (b) = e. 1-1 (b) = b 1b 2~x2xJ = b1b20 (b1 + b2) (b1 + b 2), V b = b 1b 2EB2 ·
1.e. e
~ e(00) = 00000, e(0l) = 01011
e(l0) = 10011, e(ll) = 11000.
·red group code is
Henc e thereqUl
ett(B2) ~ {e(0Q), _e(0l_), _e(10), e(ll)}
= [00000, 01011, 10011, 11000}.
:E xample 13 :
I I 0
1 0 I
0 l l
Let H =
1 O O be a parity check matrix. Determine tlie (3, 6) group code
0 I 0
0 0 I
~ : Bl ~W,
Solution:
l l 0
I O I
0 I l
Let H = 1 O O be a parity check matmc.
0 I 0
0 0 l
and ~ : IP ~ B6· be the (3, 6) encoding function Here
B3 = {cm, 100, 010, 001, 011, 101, 110, 111}
L r 1 • = • 1b 2b3 E B3 then define eH(b) = b b b ~,s,s
1 2 3

, ,._1 h ere ~ = br1 + b 2.1 + b:3.0 = b 1 + b2


JS = bl.] + bro + b3.1 = bl + b3
JS = bl.O + bi-1 + b3.1 = b2 + b3
:. ~(b) = e.i(b1 b2bJ = bl b2b3 {bl + bi) (bl + bJ (bl + bJ
~(000) = c:MXOXl, ~(001) = 001011,
eJOlO) = 010101, ~ (100} = 100110,
~(011} = 011110, ~ (101) = 101101,
~ (110) = 110011, ~(111) = 111000.
Hence, the required group code is

e.ri (IV) = {~(CXX>), eH(OlO), eH(001), e.-\(011), eH(lOO), eH(1on, eH(llO), eH(lll)}
= {<XXXXX>, 001011, 010101, 100110, 011110, 101101, 110011, 111000}

You might also like