9th International Conference on Communications and Networking in China

Research Article

Improving the intermediate recovery rate of LT codes using a novel degree distribution

  • @INPROCEEDINGS{10.4108/icst.chinacom.2014.256306,
        author={Yan Wang and Rongke Liu and Ruifeng Duan},
        title={Improving the intermediate recovery rate of LT codes using a novel degree distribution},
        proceedings={9th International Conference on Communications and Networking in China},
        publisher={IEEE},
        proceedings_a={CHINACOM},
        year={2015},
        month={1},
        keywords={lt codes; inermediate recovery rate; degree disribution; belief propagation},
        doi={10.4108/icst.chinacom.2014.256306}
    }
    
  • Yan Wang
    Rongke Liu
    Ruifeng Duan
    Year: 2015
    Improving the intermediate recovery rate of LT codes using a novel degree distribution
    CHINACOM
    IEEE
    DOI: 10.4108/icst.chinacom.2014.256306
Yan Wang1, Rongke Liu1,*, Ruifeng Duan1
  • 1: EIE school, Beihang University
*Contact email: rongke_liu@buaa.edu.cn

Abstract

The belief propagation (BP) decoding of LT codes on BEC recovers only a small subset of the source symbols when the overhead is smaller than a particular value, at which point an avalanche of decoding takes place. In this paper, a degree distribution modification method is proposed to solve this problem. The encoder adopting the modified degree distribution equals to an encoder adopting the original degree distribution which produces the encoded symbols from the original source symbols and some so-called known symbols, whose values are known to both encoder and decoder. The mean degree of the modified degree distribution is always smaller than that of the original one. Theoretical analysis gives the improvement of the intermediate recovery rate. Simulation results show that the proposed modification method is effective to Robust degree distributions and the fixed degree distribution, on both BEC and AWGN channels.