Transcript Slide 1

University of Technology
Dept. of computer Engineering and
Information Technology
Forward Error Correcting Codes for
Optical Communication
Systems
BY
Dr. Hussam Abd Ali Abdulridha
Optical Error Correcting Codes
FEC challenges in Optical
Comm. systems
-Very high information rate between 1Gb/s to
40Gb/s.
-High Coding Gain with low code rate greater than
0.7.
- Low BER between 1012 to1015 .
Optical Error Correcting Codes


FEC
Block codes

Convolutional Codes
-Binary Block Code (BCH)
 Turbo Codes
-Non-Binary Block Code (RS) (G1)


Concatenated Codes (G2)
LDPC Codes (G3)
Channel coding parameters
code generation (N,K)
Code rate = (n-k)/n
Bit redundancy (r )= (n-k)
N: codeword length
K: information length
Coding Gain (dB)= (S/N) coding (S/N) uncoding
At the same BER
First Generation Outband
FEC (Single Code)
- Reed-Solomon (RS) codes
- code-symbol interleaving of
a number of individual RS-codes
- for instance 16-way interleaved
RS(255,239) codes in the case of
ITU-T G.709
Second Generation Outband
FEC (Concatenated)
- FEC-concatenation schemes made
-Serial-concatenation of FEC codes
-recommendation ITU-T G.975.1
- super-FEC is a conc. scheme with
BCH(3860,3824) as the outer and
BCH(2040,1930) as the inner code
Product Codes
- Serially concatenated codes using two or more short
block codes to form long block codes
-If C1) n1, k1) of minimum distance dmin1 and
C2 )n2, k2) of minimum distance dmin2 are two
systematic linear block codes
n2
Coding columncolumn using C1
k2
Input Row-Row
code by C2
k1
Information
Symbol
n1
Check of column
Check
of
row
Check
of the
check
Construction of a product code
This bits
check of the
check
3rd Generation Outband FEC
(Super-FEC)
- Low-Density Parity-Check (LDPC)
coding
- leverage iterative decoding 
- very high coding-gains 

Low-Density Parity Check Codes
(LDPC)
- LDPC codes have large minimum Hamming
distance
- Parity-check matrix of a simple linear block code
(Local code) used to generate LDPC code matrix
- LDPC codes matrix depend on Local code,
codeword length, and permutation matrix
Effect of Coding on QPSK receiver with coherent
demodulation Sensitivity
0
10
UNCODED
LDPC(3969,3591)
LDPC(3969,3213)
LDPC(3969,2835)
-2
10
UNCODED
BCH(255,239)+BCH(255,239)
BCH(255,239)+BCH(255,223)
BCH(255,223)+BCH(255,223)
RS(255,239)+RS(255,239)
RS(255,239)+RS(255,223)
RS(255,223)+RS(255,223)
-2
10
-4
10
-4
10
BER
-6
10
-6
10
LDPC code gives small
increase in CG than
concat. Of RS codes at
the same code rate 0.81
-8
10
-8
10
1 dB
4.8 dB
-10
10
-10
10
-12
-12
10
-64
RS gives higher
CG than BCH at
Same code rate
-62
-60
-58
-56
-54
-52
10
-62
-60
-50
-48
Received power (dBm)
-58
-56
-54
-52
-50
-48
BER versus received power for coherent QPSK receiver operating at
1Gb/s rate and = 3.78*10-4 incorporating coherent demodulation.
Effect of Coding on QPSK receiver with
differential demodulation Sensitivity
0
10
UNCODED
LDPC(3969,3591)
LDPC(3969,3213)
LDPC(3969,2835)
-2
10
UNCODED
BCH(255,239)+BCH(255,239)
BCH(255,239)+BCH(255,223)
BCH(255,223)+BCH(255,223)
RS(255,239)+RS(255,239)
RS(255,239)+RS(255,223)
RS(255,223)+RS(255,223)
-2
10
-4
10
-4
10
LDPC code gives high CG
than concat. Of RS at the
same code rate 0.81
BER
-6
10
-8
10
-6
10
Decrease code rate in
LDPC code gives small
increase in CG
-8
10
3.6 dB
-10
2 dB
-12
-12
-62
4.3dB
-10
10
10
10
RS gives higher
than BCH at Same
code rate
10
-60
-58
-56
-54
-52
-50
-48
-46
-60
-58
-56
-54
-52
-50
-48
Received power (dBm)
BER versus
for heterodyne QPSK receiver operating at 1Gb/s rate
and PR = -46.93 dBm incorporating differential demodulation.
-46
DESIGN CONSIDERATIONS
- Processing delays: Optical communications are
particularly sensitive to delays,
- Configurable redundancy: Optical networking
applications of different range require different
levels of protection, with respect to the Quality of
Service (QoS)
- Rich statistics FEC: is not a panacea; it is 
rather introduced to obtain the necessary system
margin to guarantee QoS

