Scribd
Upload
5 views

ECL 439 WCOM g error control

Uploaded by

Gopi
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
5 views

ECL 439 WCOM g error control

Uploaded by

Gopi
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 14

ERROR CONTROL CODING

Strategies for wireless communication


Block codes
●
Hamming, Bose-Chaudhary-Hocquenghem (BCH), Reed-
Solomon (RS) codes
●
Classical and powerful, useful for wire-line, LoS satellite or
deep space communication (AWGN channel)
●
They can handle burst errors in wireless channels but not
large number of distributed errors.
●
Based on hard decision decoding, do not utilize advantages
of soft-decision decoding.
Block codes
●
Limited applications for wireless
communications.
– Large SNR wireless communication
– Outer code in a concatenated coding scheme with
convolutional code as inner code. Most errors
corrected by convolutional code, remaining errors
corrected by block code.
Classical Convolutional Encoding
Classical Convolutional Encoding
Turbo encoder
Recursive convolutional code
Convolutional encoding
●
Classical
– Non-recursive (feed forward paths only)
– Non-systematic (data bits lose their identity)
●
Turbo
– Recursive (feedback and feed-forward, both types of
paths are used)
– Systematic (data bits produced as subset of output bits)
Comparison of turbo and classical
●
Shannon’s channel capacity theorem implies that
the code-words need to be random for attaining
the target of low Pe. At the same time, the code-
words need to be decodable for practical utility.
●
Turbo codes are both, random as well as
decodable, compared to classical.
Comparison of turbo and classical
●
Turbo codes perform better for high code-rate or
low SNR channels.
●
Both need flushing of data bits from the encoder
at the end of transmission. Classical encoders
are easy to flush out but not so for Turbo codes.
Flushing is NOT done in many Turbo codes.
Comparison of turbo and classical
●
Turbo codes have error-floor. BER drops very
quickly at the beginning but later it levels off and
decreases at a much slower rate. Leveling off
happens in the range of BER = 10-5 to 10-8
depending on flushing policy, block size and type
of interleaver.
●
Both perform better with soft-decision decoding.
Comparison of turbo and classical
●
Classical codes use Viterbi decoding algorithm.
Maximum likelihood algorithm for a BSC is the
same as minimum Hamming distance.
●
Complexity of Viterbi algorithm grows
exponentially with the number of states in
encoders. Upto 256-state decoders are used in
practice.
Comparison of turbo and classical
●
Turbo codes deploy two decoders, each with 16-state or 8-
state trellis.
●
It generally deploys MAP algorithm in the iterative procedure.
●
Generally, maximum iterations are 8.
●
The combined effect is that complexity of Turbo decoding
remains manageable, yet achieving more coding gain
compared to classical codes.
Comparison of turbo and classical
●
Turbo codes are clear winner fo AWGN channel
but there is no clear winner for wireless
channels.
●
Both need inter-leavers for managing burst
errors that occur frequently in wireless
environment.

You might also like