Scribd
Upload
28 views

Turbo Codes Decoding Apps

The document discusses turbo codes and their use in iterative decoding. Turbo codes use an interleaver to increase the trellis size of the code, making direct decoding computationally infeasible. Instead, an iterative decoding process is used where two component decoders exchange soft information over multiple iterations to jointly decode the received symbols. This iterative decoding approach improves performance over traditional hard-decision decoding by up to 0.7 dB. The document outlines the basic turbo decoding algorithm and flow and provides examples of turbo code applications in wireless communications, digital video broadcasting, and deep-space telemetry standards.

Uploaded by

vasantha priyanka
Copyright
© © All Rights Reserved
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
28 views

Turbo Codes Decoding Apps

The document discusses turbo codes and their use in iterative decoding. Turbo codes use an interleaver to increase the trellis size of the code, making direct decoding computationally infeasible. Instead, an iterative decoding process is used where two component decoders exchange soft information over multiple iterations to jointly decode the received symbols. This iterative decoding approach improves performance over traditional hard-decision decoding by up to 0.7 dB. The document outlines the basic turbo decoding algorithm and flow and provides examples of turbo code applications in wireless communications, digital video broadcasting, and deep-space telemetry standards.

Uploaded by

vasantha priyanka
Copyright
© © All Rights Reserved
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
You are on page 1/ 13

Turbo Codes Decoding and

Applications
Bob Wall
EE 548

The Turbo Decoder


Since a Turbo Code trellis would have a
very large number of states (due to the
interleaver), a normal decoder would be
huge
Instead, perform an iterative decoding
process
This requires some changes to the normal
decoder instead of making a hard decision,
it needs to make a soft decision

Decoder Decisions
A hard decision is choosing a sequence of 0 or 1
output bits based on the input symbols. It does not
take into account how likely this string of bits was.
A soft decision produces a value in [0, 1] that
corresponds to the likelihood (or log likelihood).
This can be used as input to help make decisions
about the transmitted sequence

Decoder Algorithms
Normal decoding algorithms (i.e. Viterbi algorithm)
find the most likely sequence of bits that was
transmitted.
In a turbo decoder, want to find the likelihood of each
bit. This serves as the a priori probability or the
reliability of each bit, to use as input to the next
decoder
Optimal MAP (Maximum a Priori) BCJR (Bahl, Cocke,
Jelinek, Raviv)
Simpler - SOVA (Soft Output Viterbi Algorithm) lose
roughly .7 dB coding gain

Decoder Flow
Simplified flow diagram (from
http://www.xenotran.com/turbo_tech_error_turbo.html)

Block Diagram
Generic Turbo Decoder
xk
yk

DEC-

Interleaver

extrinsic information

DEC|

De-interleaver

extrinsic information

De-interleaver
xi yi

Turbo Decoding Delays


Delay introduced by interleaver
depending on how it is designed, some or
all of a block of data must be present before
encoding or decoding can begin
Delay introduced by iteration
You have to have some drawbacks no free
lunch!

Performance

Performance (cont.)

Applications
Terrestrial Wireless Communications
UMTS (Universal Mobile Telecommunication
Systems)

Digital Video Broadcasting (DVB)


CCSDS (Consultative Committee for Space
Data Systems) Telemetry Standard

CCSDS Telemetry Standard


Upgraded to specify Turbo Code as an
alternative channel coding standard
Alternative to older codes: (255, 223) ReedSolomon code, 64-state rate-1/2 convolution
coder, and their serial concatenation
2.5 dB coding gain over old standard

Also allows for symbol synchronization

Space Telemetry
Symbol Synchronization recover clock
information from the signal itself
Requires relatively frequent changes in
signal (run lengths must be limited)
Specifications suggest max run length of 64
symbols, and a minimum number of
transitions in 1000 symbols of 275

Symbol Synchronization
With turbo code specified in standard (N =
8920), simulation has shown that the max
run lengths for the rate 1/2 and rate 1/3
encoders are 42 and 53, respectively, and the
minimum number of transitions per 1000
symbols are 407 and 408, respectively
Turbo code is very close to the ideal random
code

You might also like