Upload
hilmi92
View
218
Download
0
Embed Size (px)
Citation preview
8/2/2019 Lect 6_Error Control Coding III
1/10
SET 358 3 DIG ITAL CO MMUNICATIO N
Dr. Nurul Muazzah Abdul Latif f
Faculty of Electrical Engineer ing, UTM
1
Content
O ptimum decod ing
Maximum Likelihood decoder
Soft decisions and hard decisions?
Viterbi algorithm
2
8/2/2019 Lect 6_Error Control Coding III
2/10
O ptimum decod ing
If the input seq uence messages a re e qually likely, the op timumde cod er which minimizes the prob ab ility of error is theMaximum likelihooddecoder.
ML decoder, selects a codeword among all the possiblecodewor ds which max imizes the likelihood f unctionwhere is the received seq uence a nd is one of thepossible codewords.
)( )(m
p
U|Z
Z )(mU
)(max)(ifChoose )(allover
)()( mmm pp(m)
U|ZU|ZUU
ML decoding rule:
2Lcodewords
to search!!!
3
ML decod ing for memory -less channels
Due to the indep endent channel sta tistics for memory lesschannels, the likelihood function becomes
and equivalently, the log-likelihood function becomes
The pa th metric up to time index , is called the pa rtia l pa th metric.
1 1
)(
1
)()(
21,...,...,,
)( )|()|()|,...,...,,()(21
i
n
j
m
jiji
i
m
ii
m
izzz
muzpUZpUZZZpp
iU|Z
1 1
)(
1
)()()|(log)|(log)(log)(
i
n
j
m
jiji
i
m
ii
muzpUZppm U|Z
U
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
8/2/2019 Lect 6_Error Control Coding III
3/10
Binary symmetric channels (BSC)
If is the Hamming d ista nce b etween Z and U,then
Modulator
input
1-p
p
p
1
0 0
1
)( )(mm
dd UZ,
)1log(1
log)(
)1()( )(
pL
p
pdm
ppp
nm
dLdm mnm
U
U|Z
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
Sof t and ha rd decisions
In hard decision:
The demodulator makes a firm or hard decision whetherone or zero is transmitted and provides no otherinfor mat ion for the decoder such tha t how relia ble t hedecision is.
Hence, its output is only ze ro or one (the output is
quantized only to two level) which are called hard-bits.
Decod ing based on ha rd -b its is ca lled the ha rd -
decision decoding.
8/2/2019 Lect 6_Error Control Coding III
4/10
Soft and hard decision-contd
In Sof t decision: The demodulator provides the decoder with some side
information together with the decision.
The side information provides the decoder with a measure ofconfid ence for t he decision.
The demodulator outputs which a re ca lled sof t-b its, arequantized to more than two levels.
Decoding based on soft-bits, is called the soft-decisiondecoding.
O n AW G N channels, 2 dB and on fa ding channels 6 dBgain are obtained by using soft-decoding over hard-decoding.
The Viterb i a lgorithm The Viterb i alg orithm pe rf orms Ma ximum likelihood
decoding.
It find a p at h through trellis with the larg est metric(ma ximum corr ela tion or minimum d ista nce).
It processes the demodulator outputs in an iterative manner.
At ea ch step in the trellis, it compa res the metric of all pa thsentering ea ch state, a nd keep s only the pa th with the smallestmetric, called the survivor, together with its metric.
It proceeds in the tre llis by eliminating the least likely p aths.
It red uces the de cod ing complexity to !12K
L
8/2/2019 Lect 6_Error Control Coding III
5/10
The Viterb i a lgorithm - contd
Viterbi alg orithm:
A. Do the fo llowing set up :
For a d ata block of L b its, form the tr ellis. The trellis hasL+K-1 sections or levels and sta rt s a t time and endsup at time .
La bel a ll the b ra nches in the trellis with theircorresponding branch metric.
For ea ch sta te in the tre llis a t the time which is
denoted by , def ine a pa rameterB. Then, do the f ollow ing:
it
}2,...,1,0{)(1
K
itS
ii
ttS ),(
1t
KLt
1t
The Viterb i a lgorithm - contd
1 . Set and
2 . At time , comp ute the p ar tia l p at h metrics for a ll thepaths entering each state.
3 . Set eq ua l to the b est pa rtia l pa th metricentering each sta te at time .
Keep the survivor pa th a nd delete the dea d p at hs f romthe tr ellis.
1 . If , increa se by 1 and return to step 2.
C. Start at state zero at t ime . Follow thesurviving b ra nches ba ckwards through the trellis.The pa th thus def ined is unique a nd correspond tothe M L codeword .
0),0( 1 t.2i
it
ii ttS ),(it
KLi i
KLt
8/2/2019 Lect 6_Error Control Coding III
6/10
Examp le of Hard decision Viterb idecoding
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
6t1t 2t 3t 4t 5t
)101(m)1110001011(U)0110111011(Z
Examp le of Hard decision Viterb i
decoding-contd
Label all the branches with the branch metric (Hammingdistance)
0
2
0
1
2
1
0
1
1
0
1
2
2
1
0
2
1
1
1
6t1t 2t 3t 4t 5t
1
0
ii ttS ),(
8/2/2019 Lect 6_Error Control Coding III
7/10
Examp le of Hard decision Viterb idecoding-contd
i= 2
0
2
0
1
2
1
0
1
1
0
1
2
2
1
0
2
1
1
1
6t1t 2t 3t 4t 5t
1
0 2
0
Examp le of Hard decision Viterb i
decoding-contd
i= 3
0
2
0
1
2
1
0
1
1
0
1
2
2
1
0
2
1
1
1
6t1t 2t 3t 4t 5t
1
0 2 3
0
2
30
8/2/2019 Lect 6_Error Control Coding III
8/10
Examp le of Hard decision Viterb idecoding-contd
i= 4
0
2
0
1
2
1
01
1
0
1
2
2
1
0
2
1
1
1
6t1t 2t 3t 4t 5t
1
0 2 3 0
3
2
3
0
2
30
Examp le of Hard decision Viterb i
decoding-contd
i= 5
0
2
0
1
2
1
01
1
0
1
2
2
1
0
2
1
1
1
6t1t 2t 3t 4t 5t
1
0 2 3 0 1
3
2
3
20
2
30
8/2/2019 Lect 6_Error Control Coding III
9/10
Examp le of Hard decision Viterb idecoding-contd
i= 6
0
2
0
1
2
1
01
1
0
1
2
2
1
0
2
1
1
1
6t1t 2t 3t 4t 5t
1
0 2 3 0 1 2
3
2
3
20
2
30
Exa mple of Hard decision Viterbi decoding-
contd
Trace back and then:
0
2
0
1
2
1
01
1
0
1
2
2
1
0
2
1
1
1
6t1t 2t 3t 4t 5t
1
0 2 3 0 1 2
3
2
3
20
2
30
)100( m)0000111011( U
8/2/2019 Lect 6_Error Control Coding III
10/10
Example of soft-decision Viterbi decoding
)101( m)1110001011( U
)101(m)1110001011(U
)1,3
2,1,
3
2,1,
3
2,
3
2,
3
2,
3
2,1( Z
5/3
-5/3
4/3
0
0
1/3
1/3
-1/3
-1/3
5/3
-5/3
1/3
1/3
-1/3
6t1t 2t 3t 4t 5t
-5/3
0 -5/3 -5/3 10/3 1/3 14/3
2
8/3
10/3
13/33
1/3
5/35/3
ii ttS ),(
Branch metric
Partial metric
1/3
-4/35/3
5/3
-5/3