Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Error Control Coding Text: Error Control Coding Fundamentals & Applications - Shu Lin, D.J. Costello

Download as pdf or txt
Download as pdf or txt
You are on page 1of 16

Error Control Coding

Text : Error Control Coding Fundamentals & Applications - Shu Lin, D.J. Costello

Chapters l, 2 & 3

Topics: Channel Coding ·

Review of Algebra; Groups, Rings, Fields, Vector Spaces, Matrices,

Linear Block Codes; Introduction, Syndrome & Error Detection, Minimum


distance, Error detecting & Correcting, Array & Syndrome decoding,
Probability of an Undetected error, Hamming Codes

Introduction to Cyclic Codes, Convolutional Codes

1.0 Introduction

Communication System:

x Source u Channel
Source Encoder,
Encoder
Modulation
v

Noise
Channel

Destination x̂ Source û
Decoder
x̂ Decoder Demodulation

Error Control Coding 1 Fall 21


For the purposes of this discussion the following model is sufficient.

u v
Digital
Encoder
Source

Noise
Coding
Channel

Digital û r
Decoder
Sink

* There is a duality between the problems of data compression ( Source coding - finding
the optimal source code ) and data transmission ( Channel coding). During compression
, we remove all the redundancy in the data to form the most compressed version
possible, whereas during data transmission, we add redundancy in a controlled fashion
to combat errors in the channel.

Error Control Coding 2 Fall 21


1.1 Review of Algebra:

1.11 Groups:

Definition: A set G on which a binary operation * is defined is called a group if the


following conditions are satisfied:

1. Closure For every a,b in G c= a* b is in G.

2. Associativity The binary operation is associative

i.e. for any a,b,c in G , a * ( b * c ) = (a * b ) * c

3. Identity G contains an element e such that, for any a in G

a*e = e*a = a

This element e is called the identity element of G.

4. Inverses For any element a in G, there exists another element b in G such that

a*b = b*a = e

The element b is called the inverse of a.

If G has a finite number of elements then it is called a finite group and the number of
elements in G is called the order of G.

Some groups satisfy the additional property that for all a,b in the group

a*b = b*a.

This is called the commutative property. These groups are called commutative groups or
abelian groups.

Error Control Coding 3 Fall 21


In the case of an abelian group, the symbol for the group operation is written + and is
called "addition" (even though it might not be the usual arithmetic addition).

e.g. The group G = { 1,2 ..... p-l} under modulo-p multiplication is called a
multiplicative group. If p is not a prime then the set is not a group under modulo-p
multiplication.

G={ 1,2,3,4} under modulo- 5 multiplication

. 1 2 3 4
1 1 2 3 4
2 2 4 1 3
3 3 1 4 2
4 4 3 2 1

Modulo - 5 Multiplication

* Let H be a nonempty subset of G. The subset H is said to be a subgroup of G if 11 is


closed under the group operation of G and satisfies all the conditions of a group.

e.g. Group: real numbers under addition, Sub group: integers under addition.

Error Control Coding 4 Fall 21


1.12 Rings

A ring is an abstract set that is an abelian group and also has an additional structure.

Definition: A ring R is a set with two operations defined: the first is called addition
(denoted by + ); the second is called multiplication ; and the following axioms are
satisfied.

1. R is an abelian group under addition (+).

2. Closure For any a,b in R, the product ab is in R.

3. Associative law

a(bc) = (ab) c

4. Distributive law

a(b+c) = ab + ac (b+c)a = ba + ca

The addition operation is always commutative in a ring, but the multiplication operation
need not be commutative. A commutative ring is one in which multiplication is
commutative, i.e. ab = ba for all a,b in R.

The distributive law in the definition of a ring links the addition and multiplication
operations.

e.g.
1. Thc set of all n by n matrices with integer valued elements under matrix addition and
multiplication is a noncommutative ring with identity.

2. The set of all integers under the usual addition and multiplication is a ring with
identity. This ring is conventionally denoted by Z.

Error Control Coding 5 Fall 21


1.13 Fields

Loosely speaking, an abelian group is a set in which one can add and subtract, and a
ring is a set in which one can add, subtract and multiply. A more powerful algebraic
structure, known as a field is a set in which one can add, subtract, multiply and divide.

Definition: A field F is a set that has two operations defined on it: addition and
multiplication, such that the following axioms are satisfied.

1. The set is an abelian group under addition.

2. The field is closed under multiplication, and the set of nonzero elements
is an abelian group under multiplication.

3. The distributive law

(a+b)c = ac + bc

holds for all a,b,c in the field.

It is conventional to denote the identity element under addition by 0 and to call it "zero";
to denote the additive inverse of a by -a; to denote the identity element under
multiplication by 1 and to call it "one"; and to denote the multiplicative inverse of a by
a-1.

By subtraction (a-b) we mean a + (-b); by division (a/b), we mean b-1 a.

e.g.

1. R: the set of all real numbers.


2. C: the set of complex numbers.
3. Q: the set of rational numbers.

These fields all have an infinite number of elements. We are interested in fields with a
finite number of elements.

Error Control Coding 6 Fall 21


A field with q elements, if it exists, is called a finite field or a Galois field and is
denoted by the label GF(q).

( The Galois field are named for Evarsite Galois (1811 - 1832). Abelian groups are
named for Niels Henrik Abel (1802-1829)).

What is the smallest field? It must have an element zero and an element one. In fact
these suffice with the addition and multiplication tables.

+ 0 1 . 0 1
0 0 1 0 0 0
1 1 0 1 0 1

This is the field GF(2). No other field exists with two elements.

The field GF(3)= { 0,1,2} with the operations;

+ 0 1 2 . 0 1 2
0 0 1 2 0 0 0 0
1 1 2 0 1 0 1 2
2 2 0 1 2 0 2 1

This is modulo - 3 addition and multiplication.

The field GF(4) = {0,1,2,3} with the operations;


. 0 1 2 3 . 0 1 2 3
0 0 1 2 3 0 0 0 0 0
1 1 0 3 2 1 0 1 2 3
2 2 3 0 1 2 0 2 3 1
3 3 2 1 0 3 0 3 1 2

Notice that multiplication in GF(4) is not modulo-4 and addition is not modulo-4.

Definition: Let F be a field. A subset of F is called a subfield if it is a field under the


inherited addition and multiplication. The original field F is then called an extension
field of the subfield.

e.g. GF(2) is contained in GF(4) = GF(22). However GF(2) is not contained in GF(3).

A field has all the properties of a ring. It also has an additional important property - it is
always possible to cancel. Cancellation is a weak form of division that states that

Error Control Coding 7 Fall 21


if ab = ac then b=c.

Some rings may also satisfy this cancellation law and yet not be fields. The ring of
integers is a simple example.

Definition: An integral domain is a commutative ring in which b=c whenever ab = ac


and a is nonzero.

In general, we can construct a code with symbols from any Galois field GF(q), where q
is either a prime p or a power of p. However , codes with symbols from the binary field
GF(2) or its extension GF(2m) are most widely used in digital data transmission and
storage systems because information in these systems is universally coded in binary
form for practical reasons.

Error Control Coding 8 Fall 21


1.2 Vector Spaces:

Let V be a set of elements on which a binary operation called addition + is defined. Let
F be a field. A multiplication operation, denoted by '.' , between the elements in F and
elements in V is also defined. The set V is called a vector space over the field F if it
satisfies the following conditions:

1. V is an abelian group under addition.


2. For any element a in F and any element v in V, a.v is an element in V.
3. (Distributive laws) For any elements u and v in V and any elements a
and b in F,

a.(u + v) = a .u + a .v
(a+b).v = a.v + b.v

4. (Associative law) For any v in V and any a and b in F,

(a.b).v = a.(b.v)

5. Let 1 be the unit element of F. Then for any v in V, l.v = v

The elements of V are called vectors and the elements of the field F are called scalars.
The addition on V is called a vector addition and the multiplication that combines a
scalar in F and a vector in V into a vector in V is referred to as scalar multiplication (or
product). The additive identity of V is denoted by 0.

Properties:

1. Let 0 be the zero element of the field F. For any vector v in V 0.v = 0.
2. For any scalar c in F c.0 = 0.
3. For any scalar c in F and any vector v in V,

(-c).v = c.(-v) = -(c.v)

i.e., (-c).v or c.(-v) is the additive inverse of the vector c.v.

Error Control Coding 9 Fall 21


Next a very useful vector space over GF(2) which plays a central role in coding theory.

Consider an ordered sequence of n components,

(a0 , a1,. ..., an-1)

where each component ai is all element from the binary field GF(2). This
sequence is generally called an n-tuple over GF(2).

* 2n distinct n-tuples, denote by ==> Vn

For any u Vn, u = (u0, u1, ..., un-1) and for any v Vn, v = (v0, v1, ..., vn-1)
u+v = (u0+v0, u1+v1, ..., un-1+vn-1)

* u+v is also an n-tuple over GF(2).

==> Vn is an abelian group under the addition defined above.

( check the conditions ??)

* scalar multiplication, for any a GF(2)

a. (v0, v1, ..., vn-1) = (a.v0, a.v1, ..., a.vn-1)

where a.vi – modulo-2 multiplication.

==> Vn of all n-tuples over GF(2) forms a vector space over GF(2).

E.g. n=5 consists of 32, 5-tuples over GF(2).

Theorem: Let S be a nonempty subset of a vector space V over a field F. Then S is a


subspace of V if the following conditions are satisfied.

1. For any u,v in S, u+v is also in S.


2. For any a in F and any u in S, a.u is also in S.

Error Control Coding 10 Fall 21


e.g. V5

{(0 0 0 0 0), (0 0 1 1 1), (1 1 0 1 0), (1 1 1 0 1)}

Let v1, v2, ..., vk be k vectors in a vector space V over a field. Let a1, a2, ..., ak be k
scalars from F. The sum

a1v1+ a2v2+ ...+ akvk

is called a linear combination of v1, v2, ..., vk. Clearly,

(a1v1+ a2v2+ ...+ akvk) + (b1v1+ b2v2+ ...+ bkvk)


= (a1+ b1) v1+ (a2+ b2)v2+ ...+ (ak+ bk)vk

is also a linear combination of v1, v2, ..., vk and the product of a scalar c in F and a linear
combination of v1, v2, ..., vk

c.(a1v1+ a2v2+ ...,+ akvk) = (ca1)v1+ (ca2)v2+ ...+ (cak)vk

is also a linear combination of v1, v2, ..., vk. It follows from the above theorem,

Theorem: Let v1, v2, ..., vk be k vectors in a vector space V over a field of F. The set of
all linear combinations of v1, v2, ..., vk forms a subspace of V.

e.g. V5 the linear combinations of (0 0 1 1 1), (1 1 1 0 1).

Error Control Coding 11 Fall 21


A set of vectors v1, v2, ..., vk in a vector space V over a field F is said to be linearly
dependent if and only if there exists k scalars a1, a2, ..., ak, from F, not all zero, such that

a1v1+ a2v2+ ...+ akvk = 0

A set of vectors v1, v2, ..., vk is said to be linearly independent if it is not linearly
dependent, i.e. if v1, v2, ..., vk are linearly independent, then

a1v1+ a2v2+ ...+ akvk  0

Unless a1=a2= ...= ak = 0.

e.g. The vectors (1 0 1 1 0), (0 1 0 01) and (1 1 1 1 1)are linearly dependent since

1.(1 0 1 1 0) + 1.(0 1 0 0 1) + l.(l 1 1 11) = (0 0 0 0 0).

However, (1 0 1 1 0),(0 1 0 0 1) and (1 1 0 1 1)are linearly independent. All eight


possible linear combinations would not give 0.

A set of vectors is said to span a vector space V if every vector if V is a linear


combination of vectors in the set. In any vector space or subspace there exists at least
one set B of linearly independent vectors which span the space. This set is called a basis
(or base) of the vector space. The number of vectors in a basis of a vector space is
called the dimension of the vector space. ( Note that the number of vectors in any two
bases are thc same).

Error Control Coding 12 Fall 21


Consider the vector space Vn of all n-tuples over GF(2). Let us form the following n n-
tuples:

e0 = (1 , 0, 0 ......... 0)

e1 = (0, 1, 0,. .......... 0)

e2 = (0, 0, 1, 0 ........ 0)
.
.
en-1 = (0, 0, 0,.. . . ., l)

where the n-tuple ei has only one nonzero component at I th position. Then, every n-
tuple (a0, ..., an-1), in Vn can be expressed as a linear combination of e0, ..., en-1 , as
follows:

(a0, ..., an-1) = a0e0 + a1e1 +.......+ an-1en-1

Therefore, e0, ..., en-1, span the vector space Vn of all n-tuples over GF(2). It is clear that
e0, ..., en-1 are linearly independent. Hence they form a basis for V, and the dimension is
n. If k< n and v1, ..., vk are k linearly independent vectors in V. the all linear
combinations of v1, ..., vk of the form

u = c1v2 + c2v2 +.......+ ckvk

form a k-dimensional subspace S of V,. Since each ci has two possible values 0 or 1,
there are 2k possible distinct linear combinations of v1, ..., vk. Thus S consists of 2k
vectors and is a k-dimensional subspace of Vn.

Error Control Coding 13 Fall 21


Let u = (u0, u1, ..., un-1) and v = (v0, v1, ..., vn-1) be any two n-tuples in Vn.

We define the inner product (or dot product ) of u and v as

u.v = u0.v0 + u1.v1, ...+ un-1.vn-1

where ui.vi and ui.vi + ui+1.vi+1 are carried out in modulo-2 multiplication and addition.
Hence the inner product u.v is a scalar in GF(2). If u.v = 0, u and v are said to be
orthogonal to each other. The inner product has the following properties.

i) u.v = v.u
ii) u.(v + w ) = u.v + u.w
iii) (au).v = a. ( u.v )

The concept of inner product can be generalized to any Galois field.

Let S be a k-dimensional subspace of Vn and let Sd be the set of vectors in Vn such that,
for any u in S and v in Sd u.v = 0. The set Sd contains at least the all-zero n-tuple 0 =
(0,0,.,0) since for any u in S 0.u = 0.Thus Sd is nonempty. For any element a in GF(2)
and any v in Sd.

a.v = 0 if a=0
=v if a=1

Therefore a.v is also in Sd. Let v and w be any two vectors in Sd. For any vector u in S,
u.(v+w) = u.v + u.w = 0+0=0. This says that if v and w are orthogonal to u the vector
sum v + w is also orthogonal to u.

==> v + w is a vector in Sd. ==> Sd is also a subspace of Vn.

This subspace Sd is called the null ( or dual) space of S. Conversely S is also the null
space of Sd.

Error Control Coding 14 Fall 21


Theorem: Let S be a k-dimensional space of the vector space V. of all n-tuples over
GF(2). The dimension of its null space Sd is n - k.

i.e., dim(S) + dim(Sd ) = n.

1.3 Matrices

A k x n matrix over GF(2) ( or over ally other field) is a rectangular array with k rows
and n columns.

g 00 g 01 .......g 0 ,n −1 
 
g 10 g 11 .......g 1,n −1 
G = . 
 
. 
g g k −1,1 .......g k −1,n −1 
 k −1,0

where each entry g ii with 0< i<k and 0< j<n is an element from the binary field GF(2).

* each row of G is an n-tuple over GF(2) and each column is a k-tuple over GF(2). G
can also be represented its

g 0 
g 
 1 
G = . 
 
. 
g k −1 

If the k (k < n) rows of G are linearly independent then the 2k linear combinations of
these rows from a k-dimensional subspace of the vector space V. of all n-tuples over
GF(2). This subspace is called the row space of G.

We may interchange any two rows of G or add one row to the another. These are called
elementary row operations. Performing elementary row operations on G we obtain
another matrix G' over GF(2). However both G and G' have the same row space.

Error Control Coding 15 Fall 21


Let S be the row space of a kxn matrix G over GF(2) whose k rows g0, g1, ..., gk-1, are
linearly independent. Let Sd be the null space of S. Then the dimension of Sd is n – k.
Let h0, h1, ..., hn-k-1 be n-k linearly independent vectors in Sd.

We may form an (n - k)xn matrix H using , h0, h1, ..., hn-k-1 as rows.

h 0 
h 
 1 

H= . 
 
. 
h n −k −1 

The row space of H is Sd. Since each row gi of G is a vector in S and each row hj of H is
a vector in Sd, the inner product of gi and hj must be zero. ( i.e. gi . hj= 0). Since the row
space S of G is the null space of the row space Sd of H we call S the null space of H.

Theorem: For any kxn matrix G over GF(2) with k linearly independent rows there
exists an (n-k)×n matrix H over GF(2) with n-k linearly independent rows such that for
any row gi in G and any hj in H, gi . hj = 0. The row space of G is the null space of H
and vice versa.

e.g.

1 1 0 1 1 0  1 0 1 1 0 0
G = 0 0 1 1 1 0  , H = 0 1 1 0 1 0 
   
0 1 0 0 1 1 1 1 0 0 0 1

we can easily check that each row of G is orthogonal to each row of H.

* A kxk matrix is called tin identity matrix if it has 1's on the main diagonal and 0's
elsewhere. This matrix is usually denoted by Ik.
* A submatix of a matrix G is a matrix that is obtained by striking out given rows or
columns of G.

Error Control Coding 16 Fall 21

You might also like