DocumentCode :
918910
Title :
Raptor codes on binary memoryless symmetric channels
Author :
Etesami, Omid ; Shokrollahi, Amin
Author_Institution :
Comput. Sci. Div., Univ. of California, Berkeley, CA, USA
Volume :
52
Issue :
5
fYear :
2006
fDate :
5/1/2006 12:00:00 AM
Firstpage :
2033
Lastpage :
2051
Abstract :
In this paper, we will investigate the performance of Raptor codes on arbitrary binary input memoryless symmetric channels (BIMSCs). In doing so, we generalize some of the results that were proved before for the erasure channel. We will generalize the stability condition to the class of Raptor codes. This generalization gives a lower bound on the fraction of output nodes of degree 2 of a Raptor code if the error probability of the belief-propagation decoder converges to zero. Using information-theoretic arguments, we will show that if a sequence of output degree distributions is to achieve the capacity of the underlying channel, then the fraction of nodes of degree 2 in these degree distributions has to converge to a certain quantity depending on the channel. For the class of erasure channels this quantity is independent of the erasure probability of the channel, but for many other classes of BIMSCs, this fraction depends on the particular channel chosen. This result has implications on the "universality" of Raptor codes for classes other than the class of erasure channels, in a sense that will be made more precise in the paper. We will also investigate the performance of specific Raptor codes which are optimized using a more exact version of the Gaussian approximation technique.
Keywords :
Gaussian channels; approximation theory; binary codes; channel capacity; channel coding; convergence of numerical methods; decoding; error statistics; memoryless systems; BIMSC; Gaussian approximation technique; Raptor code; arbitrary binary input memoryless symmetric channel; belief-propagation decoder; convergence; degree distribution; error probability; information-theoretic argument; optimization; underlying channel capacity; Algorithm design and analysis; Bipartite graph; Computer science; Error probability; Gaussian approximation; Information theory; Iterative algorithms; Iterative decoding; Parity check codes; Stability; Belief-propagation; LT-codes; graphical codes; raptor codes;
fLanguage :
English
Journal_Title :
Information Theory, IEEE Transactions on
Publisher :
ieee
ISSN :
0018-9448
Type :
jour
DOI :
10.1109/TIT.2006.872855
Filename :
1624639
Link To Document :
بازگشت