Contents

1 Introduction 2

2 Error Detection 3
2.1 Eectiveness of Error-Detection Codes . . . . . . . . . . . . . 5
2.1.1 1. Random Error Vector Model . . . . . . . . . . . . . 5
2.1.2 2. Random Bit Error Model . . . . . . . . . . . . . . . 5
2.1.3 3. Burst Errors . . . . . . . . . . . . . . . . . . . . . . 6

1
1 Introduction
In most communication channels there would be a probability of bit errors
in transmission. These bit rate errors are usually very small, for example for
modern bre optic transmission channels it is about 10−9 th order. However,
in some cases it would be as high as 10−3 .The minimum or the standard
amount of bit rate error depends on the particular application of the com-
munication transmission.
In order to improve the performance of bit rate errors there are two main
error-control approaches:

ˆ The rst approach involves the detection of errors and an automatic

retransmission request (ARQ).

ˆ The second approach, forward error correction(FEC) involves the de-

tection of errors followed by further processing of the received informa-
tion that attempts to correct the errors.

The rst approach assumes the availability of a retun channel over which
the retransmission request can be made. For example, it is widely used in
computer communication systems that use telephone lines. It is also widley
used to provide reliable data transmission over the Internet.
The second approach, FEC is appropriate when a return channel is not
available, retransmission requests are not easliy accommodated, or a large
amount of data is sent and a retransmission to correct a few errors is very
inecient. FEC is used in satellite and deep-space communications.
The dierence between these two approaches is that ARQ wastes band-
width by using retransmissions, whereas FEC in general requires additional
redundancy in the transmitted information and incurs signicant processing
complexity in performing the error correction.

2
2 Error Detection

Figure 1: General error-detection system

As shown in the Figure 01 the information produced by an application

is encoded so that the stream that is input into the communication channel
satises a specic pattern or condition. The receiver checks the stream com-
ing out of the communication channel to see whether the pattern is satised.
If it does not satisfy the receiver can be certain that an error has occured
and therefore sets an alaram to alert the user. This certainty stems from
the fact that no such pattern would have been trasnmitted by the
encoder.
The simplest code is the single parity check code that takes k infor-
mation bits and appends a single check bit to form a codeword.
For example, let's consider the procedure for an even parity error detec-
tion.
Here, the parity check ensures that the total number of 1s in the codeword
is even. The check bit in this case is called a parity bit. This code is an
example of the so-called linear codes because the parity bit is calculated as
the modulo 2 sum of the information bits:
bk+1 = b1 + b2 + · · · + bk modulo2

where b1 , b2 , . . . , bk are the information bits.

In modulo 2 arithmetic: 0 + 0 = 0
0+1=1
1+0=1
1+1=0
Thus, if the information bits contain an even number of 1s, then the
parity bit will be 0: and if they contain an odd number, then the parity bit
will be 1. Consequently, this will produce a codeword that always contains
an even number of 1 s. This paatern denes the single parity check code.
In general, if the codeword udergoes an odd number errors, the corre-
sponding output block will also contain an odd number of 1s. Therefore,
the single parity bit allows us to detect all error patterns that introduce an
odd number of errors. On the other hand, the single parity bit will fail to

3
detect any error patterns that introduce an even number of errors, since the
resulting binary vector will have even parity.
An alternative method of looking at the opertaion of this exapmle(even
parity error detection) is shown in Figure 02.

Figure 2: Error Detection system using check bits.

At the transmitter a checksum is calculated from the information bits

and transmitted along with the information. At the reciver, the checksum is
recalculated, based on the received information. The recived and recalculated
checksums are locompared, and the error alarm is set if they disagree.
This example can be used to present two fundamental observations about
error detection.

1. That error detection requires redundancy in that the amount of in-

formation that is transmitted is over and above the required minimum.

For a single parity check code of length k + 1 , k bits are information

bits and one bit is the parity bit. Therefore, the fraction 1/(k + 1) of the
transmitted bits is redundant.

1. The second observation is that every error-detection technique

will fail to detect some errors. Particularly, when a valid codeword
is converted into another valid codeword due to transmission errors.

For the single parity check code, an even number of transmission errors
will always convert a valid codeword to another valid codeword.

4
2.1 Eectiveness of Error-Detection Codes
The eectiveness of an error-detection code is measured by the probability
that the system fails to detect an error.
Now, inorder to calculate this probability of error-detection failure, it is
required to know the probability of occuring various errors, which in return
depend in the particular properties of the given communication channel.
For this purpose three models of error channels are introduced:
Suppose we transmit a codeword that has n bits. Dene the error vector
e = (e1 , e2 , . . . , en ) where ei = 1 if an error occurs in the ith transmitted bit
and ei = 0 otherwise.

2.1.1 1. Random Error Vector Model

In this model, all 2n possible error vectors are equally likely to occur. In this
channel model the probability of e does not depend on the number of errors
it contains. Thus the error vector (1, 0, . . . , 0) has the same probability of
occurrence as the error vector (1, 1, . . . , 1).
The single parity check code will fail when the error vector has an even
number of 1s. Thus for the random error vector channel model, the proba-
bility of error detection failure is 1/2.

2.1.2 2. Random Bit Error Model

In this model, the bit errors occur independently of each other. A typical
example is Satellite communication.
Let p be the probability of an error in a single-bit transmission.T he
probability of an error vector that has j errors is pj (1−p)n−j ,since each of the
j errors occurs with probability p and each of the n − j correct transmissions
occurs with probability 1 − p. By rewriting this probability we obtain:

 w(e)
n−w(e) w(e) n p
p[e] = (1 − p) p = (1 − p)
1−p
where the weight w(e) is dened as the number of 1s in e . For any useful
communication channel, the probability of bit error is much smaller than 1,
and so p < 1/2 and p/(1 − p) < 1. This implies that for the random bit error
channel the probability of e decreases with the weight w(e), that is, as the
number of errors (1s) increases. In other words, an error pattern with given
number of bit errors is more likely than an error pattern with a larger number

5
of bit errors. Therefore this channel tends to map a transmitted codeword
into binary blocks that are clustered around the codeword.
The single parity check code will fail if the error pattern has even number
of 1s. Therefore, in the random bit error model:

P [error detection f ailure] = P [undetectable error pattern]

= P [error patterns with even number of 1s]
   
n 2 n−2 n 4
= p (1 − p) + p (1 − p)n−4 + . . .
2 4

2.1.3 3. Burst Errors