On the M*-BCJR and LISS Algorithms for Soft-Output Decoding of Convolutional Codes
(2006)Department of Electrical and Information Technology
- Abstract (Swedish)
- To communicate reliably over noisy wireline or wireless channels, algorithms that add redundancy are used to help reconstruct the transmitted sequence at the receivers. Such error control coding algorithms are today used in all digital communication systems. Optimum decoding algorithms such as the Viterbi and BCJR have very good performances but often too high complexity.
Therefore, less complex suboptimum algorithms are of interest.
In this thesis, two suboptimum algorithms for soft output decoding, M*-BCJR and LISS, are evaluated. Their behavior in terms of bit error rate, quality of L-values, and complexity are analyzed. An Additive White Gaussian Noise (AWGN) channel is considered. M*-BCJR and LISS are based on two different... (More) - To communicate reliably over noisy wireline or wireless channels, algorithms that add redundancy are used to help reconstruct the transmitted sequence at the receivers. Such error control coding algorithms are today used in all digital communication systems. Optimum decoding algorithms such as the Viterbi and BCJR have very good performances but often too high complexity.
Therefore, less complex suboptimum algorithms are of interest.
In this thesis, two suboptimum algorithms for soft output decoding, M*-BCJR and LISS, are evaluated. Their behavior in terms of bit error rate, quality of L-values, and complexity are analyzed. An Additive White Gaussian Noise (AWGN) channel is considered. M*-BCJR and LISS are based on two different decoding strategies, trellis based decoding and sequential decoding,
respectively. The simulation results show that M*-BCJR performs better than LISS in terms of bit error rate, quality of L-values, and complexity. When decoding a stand alone code both algorithms perform well, but in an iterative decoding scheme for serially concatenated convolutional codes, the results are less satisfying.
The soft outputs of LISS show a strong linear tendency (with depth) which distorts the quality of the L-values. To tackle this problem a variation of the LISS algorithm is proposed, LISSOM, in which certain states are merged. The LISSOM decreases the complexity but keeps the bit error rate at the same level for stand alone decoding. However, in an iterative decoding scheme the results are improved compared to LISS but still not satisfying. A suggestion for future work is to further develop the idea behind LISSOM. This may include an investigation of the metric used in LISS aiming to improve the quality of L-values. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/2117484
- author
- Larsson, Jacob and Wesslén, Fredrik
- supervisor
-
- Rolf Johannesson LU
- Maja Loncar LU
- organization
- year
- 2006
- type
- H2 - Master's Degree (Two Years)
- subject
- language
- English
- id
- 2117484
- date added to LUP
- 2011-08-29 10:09:13
- date last changed
- 2013-06-17 10:31:55
@misc{2117484,
abstract = {{To communicate reliably over noisy wireline or wireless channels, algorithms that add redundancy are used to help reconstruct the transmitted sequence at the receivers. Such error control coding algorithms are today used in all digital communication systems. Optimum decoding algorithms such as the Viterbi and BCJR have very good performances but often too high complexity.
Therefore, less complex suboptimum algorithms are of interest.
In this thesis, two suboptimum algorithms for soft output decoding, M*-BCJR and LISS, are evaluated. Their behavior in terms of bit error rate, quality of L-values, and complexity are analyzed. An Additive White Gaussian Noise (AWGN) channel is considered. M*-BCJR and LISS are based on two different decoding strategies, trellis based decoding and sequential decoding,
respectively. The simulation results show that M*-BCJR performs better than LISS in terms of bit error rate, quality of L-values, and complexity. When decoding a stand alone code both algorithms perform well, but in an iterative decoding scheme for serially concatenated convolutional codes, the results are less satisfying.
The soft outputs of LISS show a strong linear tendency (with depth) which distorts the quality of the L-values. To tackle this problem a variation of the LISS algorithm is proposed, LISSOM, in which certain states are merged. The LISSOM decreases the complexity but keeps the bit error rate at the same level for stand alone decoding. However, in an iterative decoding scheme the results are improved compared to LISS but still not satisfying. A suggestion for future work is to further develop the idea behind LISSOM. This may include an investigation of the metric used in LISS aiming to improve the quality of L-values.}},
author = {{Larsson, Jacob and Wesslén, Fredrik}},
language = {{eng}},
note = {{Student Paper}},
title = {{On the M*-BCJR and LISS Algorithms for Soft-Output Decoding of Convolutional Codes}},
year = {{2006}},
}