• DocumentCode
    897011
  • Title

    Structured low-density parity-check codes

  • Author

    Moura, José M F ; Lu, Jin ; Zhang, Haotian

  • Author_Institution
    Dept. of Electr. & Comput. Eng., Carnegie Mellon Univ., Pittsburgh, PA, USA
  • Volume
    21
  • Issue
    1
  • fYear
    2004
  • Firstpage
    42
  • Lastpage
    55
  • Abstract
    This article describes the different methods to design regular low density parity-check (LDPC) codes with large girth. In graph terms, this corresponds to designing bipartite undirected regular graphs with large girth. Large girth speeds the convergence of iterative decoding and improves the performance at least in the high SNR range, by slowing down the onsetting of the error floor. We reviewed several existing constructions from exhaustive search to highly structured designs based on Euclidean and projective finite geometries and combinatorial designs. We describe GB and TS LDPC codes and compared the BER performance with large girth to the BER performance of random codes. These studies confirm that in the high SNR regime these codes with high girth exhibit better BER performance. The regularity of the codes provides additional advantages that we did not explore in this article like the simplicity of their hardware implementation and fast encoding.
  • Keywords
    error statistics; graph theory; iterative decoding; parity check codes; random codes; BER performance; Euclidean geometries; LDPC codes; SNR regime; bipartite undirected regular graphs; bit error rate; combinatorial designs; graph terms; iterative decoding; large girth; low-density parity-check codes; projective finite geometries; random codes; signal-to-noise ratio; Bipartite graph; Bit error rate; Block codes; Equations; Graph theory; Linear code; Parity check codes; Sparse matrices; Turbo codes; Vectors;
  • fLanguage
    English
  • Journal_Title
    Signal Processing Magazine, IEEE
  • Publisher
    ieee
  • ISSN
    1053-5888
  • Type

    jour

  • DOI
    10.1109/MSP.2004.1267048
  • Filename
    1267048