QC LDPC码低复杂度消环算法

Low-complexity Cycle Elimination Algorithms for QC LDPC Codes

  • 摘要: QC LDPC (Quasi-才yclic Low-density Parity-check)是一类半结构化的低密度奇偶校验码,其分块的矩阵结构具有超大规模集成电路实现上的便利,同时保持了优异的纠错性能. 本文针对QC LDPC码的基矩阵,提出一种移位因子的搜索方法及其改进版本。通过对基矩阵的扩展矩阵的Tanner图进行树形展开来进行环的检验,避免了传统算法中的复杂算术操作,降低了复杂度。在采用和IEEE 802.16e中码率为0.5的LDPC码方案相同的基矩阵条件下,本文的算法构造出的QC LDPC码具有更优的环长分布,同时纠错性能也有提升。

     

    Abstract: Quasi-cyclic LDPC code is a kind of half-structured low density parity check code. Its block-based property leads to highly structured integrated circuits implementation, without significant performance loss. An algorithm searching shift values for the nonzero entries in base matrix of a Quasi-cyclic LDPC code is proposed along with its improved version. By expanding the Tanner graph of the corresponding expansion matrix, cycles with given length can be detected. Unlike some former algorithms, arithmetic operations are avoided, which leads to a considerable complexity reduction. Given the same base matrix with code rate 0.5 as is used in the IEEE 802.16e standard, LDPC codes constructed by the proposed algorithms have better cycle distributions and achieve slightly improved error-correction performance than the IEEE 802.16e LDPC code as well.

     

/

返回文章
返回