Convolutional Coding pot

Convolutional Coding
Today, we are going to talk about:

Another class of linear codes, known as Convolutional
codes.

Structures of the encoder and different ways for
representing it.: state diagram and trellis representation
of the code.

What is a Maximum likelihood decoder?

How the decoding is performed for Convolutional codes
(the Viterbi algorithm) ?

Soft decisions vs. hard decisions
2
Convolutional codes

Convolutional codes offer an approach to error control
coding substantially different from that of block codes.

A convolutional encoder:

encodes the entire data stream, into a single codeword.

does not need to segment the data stream into blocks of fixed
size (
Convolutional codes are often forced to block structure by periodic
truncation
).


is a machine with memory.

This fundamental difference in approach imparts a different
nature to the design and evaluation of the code.

Block codes are based on algebraic/combinatorial
techniques.

Convolutional codes are based on construction techniques.
3
Convolutional codes-cont’d

A Convolutional code is specified by three parameters
or where

is the coding rate, determining the number of data
bits per coded bit.

In practice, typically k=1 is chosen and we assume that from
now on.

K is the constraint length of the encoder a where the encoder
has K-1 memory elements.

There is different definitions in literatures for constraint length.
),,( Kkn
),/( Knk
nkR
c

/=
4
Block diagram of the DCS
Information
source
Rate 1/n
Conv. encoder
Modulator
Information
sink
Rate 1/n
Conv. decoder
Demodulator
  
sequenceInput
21
, ), ,,(
i
mmm=m
  
  
bits) coded ( rdBranch wo
1
sequence Codeword
321

, ), ,,,(
n
nijiii
i

, ,u, ,uuU
UUUU
=
=
= G(m)U
, )
ˆ
, ,
ˆ
,
ˆ
(
ˆ
21 i
mmm=m

  
  
dBranch worper outputs
1
dBranch worfor
outputsr Demodulato
sequence received
321

, ), ,,,(
n
nijii
i
i

i
, ,z, ,zzZ
ZZZZ
=
=Z
C
h
a
n
n
e
l
5
A Rate ½ Convolutional encoder

Convolutional encoder (rate ½, K=3)

3 shift-registers where the first one takes the
incoming data bit and the rest, form the memory
of the encoder.
Input data bits Output coded bits
m
1
u
2
u
First coded bit
Second coded bit
21
,uu

(Branch word)
6
A Rate ½ Convolutional encoder
1 0 0
1
t
1
u
2
u
11
21
uu
0 1 0
2
t
1
u
2
u
01
21
uu
1 0 1
3
t
1
u
2
u

00
21
uu
0 1 0
4
t
1
u
2
u
01
21
uu
)101(=m
Time
Output OutputTime
Message sequence:
(Branch word) (Branch word)
7
A Rate ½ Convolutional encoder
Encoder)101(=m
)1110001011(=U
0 0 1
5
t
1
u
2
u
11

21
uu
0 0 0
6
t
1
u
2
u
00
21
uu
Time
Output
Time
Output
(Branch word) (Branch word)
8
Effective code rate

Initialize the memory before encoding the first bit (all-
zero)

Clear out the memory after encoding the last bit (all-
zero)

Hence, a tail of zero-bits is appended to data bits.

Effective code rate :


L is the number of data bits and k=1 and Rc=1/n is assumed:
data
Encoder
codewordtail
ceff
R
KLn
L
R <
−+
=
)1(
9
Encoder representation

Vector representation:

We define n binary vector with K elements (one
vector for each modulo-2 adder). The i:th element
in each vector, is “1” if the i:th stage in the shift
register is connected to the corresponding modulo-
2 adder, and “0” otherwise.

Example:
m
1
u
2
u
21

uu
)101(
)111(
2
1
=
=
g
g
10
Encoder representation – cont’d

Impulse response representaiton:

The response of encoder to a single “one” bit that
goes through it.

Example:
11001
01010
11100
111011 :sequenceOutput
001 :sequenceInput
21
uu
Branch word
Register
contents
1110001011
1110111

0000000
1110111
OutputInput m
Modulo-2 sum:
11
Encoder representation – cont’d

Polynomial representation:

We define n generator polynomials, one for each
modulo-2 adder. Each polynomial is of degree K-1 or
less and describes the connection of the shift
registers to the corresponding modulo-2 adder.

Example:
The output sequence is found as follows:
22)2(
2
)2(
1
)2(
02
22)1(
2
)1(
1
)1(
01
1 )(
1 )(

XXgXggX
XXXgXggX
+=++=
++=++=
g
g
)()( with interlaced )()()(
21
XXXXX gmgmU =
12
Encoder representation –cont’d
In more details:
1110001011
)1,1()0,1()0,0()0,1()1,1()(
.0.0.01)()(
.01)()(
1)1)(1()()(
1)1)(1()()(
432
432
2
432
1
422
2
4322
1
=
++++=
++++=

++++=
+=++=
+++=+++=
U
U
gm
gm
gm
gm
XXXXX
XXXXXX
XXXXXX
XXXXX
XXXXXXXX
13
State diagram

A finite-state machine only encounters a finite number
of states.

State of a machine: the smallest amount of
information that, together with a current input to the
machine, can predict the output of the machine.

In a Convolutional encoder, the state is represented
by the content of the memory.

Hence, there are states.
1
2

−K
14
State diagram – cont’d

A state diagram is a way to represent the encoder.

A state diagram contains all the states and all possible
transitions between them.

Only two transitions initiating from a state

Only two transitions ending up in a state
15
State diagram – cont’d
10 01
00
11
outputNext
state
inputCurrent
state
101
010
11
011
100
10
001
110
01

111
000
00
0
S
1
S
2
S
3
S
0
S
2
S
0
S
2
S
1
S
3
S
3
S
1
S
0
S
1

S
2
S
3
S
1/11
1/00
1/01
1/10
0/11
0/00
0/01
0/10
Input
Output
(Branch word)
16
Trellis – cont’d

Trellis diagram is an extension of the state
diagram that shows the passage of time.

Example of a section of trellis for the rate ½ code
Time
i
t
1+i
t
State
00

0
=S
01
1
=S
10
2
=S
11
3
=S
0/00
1/10
0/11
0/10
0/01
1/11
1/01
1/00
17
Trellis –cont’d

A trellis diagram for the example code
0/11
0/10
0/01
1/11
1/01
1/00
0/00

0/11
0/10
0/01
1/11
1/01
1/00
0/00
0/11
0/10
0/01
1/11
1/01
1/00
0/00
0/11
0/10
0/01
1/11
1/01
1/00
0/00
0/11
0/10
0/01
1/11
1/01
1/00
0/00
6
t

1
t
2
t
3
t
4
t
5
t
1 0 1 0 0
11 10 00 10 11
Input bits
Output bits
Tail bits
18
Trellis – cont’d
1/11
0/00
0/10
1/11
1/01
0/00
0/11
0/10
0/01
1/11
1/01
1/00
0/00

0/11
0/10
0/01
0/00
0/11
0/00
6
t
1
t
2
t
3
t
4
t
5
t
1 0 1 0 0
11 10 00 10 11
Input bits
Output bits
Tail bits
19
Block diagram of the DCS
Information
source
Rate 1/n
Conv. encoder
Modulator

Information
sink
Rate 1/n
Conv. decoder
Demodulator
  
sequenceInput
21
, ), ,,(
i
mmm=m
  
  
bits) coded ( rdBranch wo
1
sequence Codeword
321

, ), ,,,(
n
nijiii
i
, ,u, ,uuU
UUUU
=
=
= G(m)U
, )
ˆ
, ,

ˆ
,
ˆ
(
ˆ
21 i
mmm=m

  
  
dBranch worper outputs
1
dBranch worfor
outputsr Demodulato
sequence received
321

, ), ,,,(
n
nijii
i
i
i
, ,z, ,zzZ
ZZZZ
=
=Z
C
h
a

n
n
e
l
20
Optimum decoding

If the input sequence messages are equally likely,
the optimum decoder which minimizes the probability
of error is the Maximum likelihood decoder.

ML decoder, selects a codeword among all the
possible codewords which maximizes the likelihood
function where is the received
sequence and is one of the possible codewords:
)(
)(m
p
′
U|Z
Z
)(m
′
U
)(max)( if Choose
)(
allover
)()( mmm
pp
(m)

U|ZU|ZU
U
=
′′

ML decoding rule:

codewords
to search!!!
L
2
21
ML decoding for memory-less channels

Due to the independent channel statistics for
memoryless channels, the likelihood function becomes
and equivalently, the log-likelihood function becomes

The path metric up to time index , is called the partial path
metric.
∏∏∏
∞
= =
∞
=
===
1 1
)(
1
)()(

21, , ,,
)(
)|()|()|, , ,,()(
21
i
n
j
m
jiji
i
m
ii
m
izzz
m
uzpUZpUZZZpp
i
U|Z
∑∑∑
∞
= =
∞
=
===
1 1
)(
1
)()(
)|(log)|(log)(log)(
i

n
j
m
jiji
i
m
ii
m
uzpUZppm U|Z
U
γ
Path metric Branch metric
Bit metric

ML decoding rule:
Choose the path with maximum metric among
all the paths in the trellis.
This path is the “closest” path to the transmitted sequence.
""i
22
Binary symmetric channels (BSC)

If is the Hamming distance between Z
and U, then

Modulator
input
1-p
p
p

1
0 0
1
)(
)(m
m
dd UZ,=
)1log(
1
log)(
)1()(
)(
pL
p
p
dm
ppp
nm
dLd
m
mnm
−+









−
−=
−=
−
U
U|Z
γ
Demodulator
output
)0|0()1|1(1
)1|0()0|1(
ppp
ppp
==−
==

ML decoding rule:
Choose the path with minimum Hamming distance
from the received sequence.
Size of coded sequence
23
AWGN channels

For BPSK modulation the transmitted sequence
corresponding to the codeword is denoted by
where and
and .

The log-likelihood function becomes


Maximizing the correlation is equivalent to minimizing the
Euclidean distance.
)(m
U
cij
Es ±=
>=<=
∑∑
∞
= =
)(
1 1
)(
)(
m
i
n
j
m
jiji
szm SZ,
U
γ
Inner product or correlation
between Z and S

ML decoding rule:
Choose the path which with minimum Euclidean distance
to the received sequence.
), ,, ,(

)()()(
1
)(
m
ni
m
ji
m
i
m
i
sssS =
, ), ,,(
)()(
2
)(
1
)( m
i
mmm
SSS=S
24
Soft and hard decisions

In hard decision:

The demodulator makes a firm or hard decision
whether one or zero is transmitted and provides no
other information for the decoder such that how
reliable the decision is.


Hence, its output is only zero or one (the output is
quantized only to two level) which are called “hard-
bits”.

Decoding based on hard-bits is called the
“hard-decision decoding”.
25