LDPC一致校验矩阵的LU分解算法

LU Decomposition Algorithm of LDPC’s Unanimous Checking Matrix

不同于Turbo码,LDPC是数学结构非常好的一类高速码。对其结构进行系统研究,不但对实际编码方案很有帮助,而且也有很高的理论价值。二进制低密度码(MN码)编码的基础为Gaussian消元法(消去法)。文中推导了MN码的一致校验矩阵的基于初等变换的Gaussian消元法。研究表明,基于初等变换的Gaussian消元法结合"minprod"算法在进行LU分解时,可以有效地保持MN码一致校验矩阵的稀疏性。

The basis of Binary Low-Density Parity-Check Code is Gaussian Elimination, so Gaussian elimination algorithm of MNCode＇ s unanimous checking matrix based on elementary transform is deduced. The research results indicate that, when decomposing the check matrix the Gaussian Elimination based on Elementary Transform, by using ＂minprod＂ -Algorithm, can effectively keep the sparseness of unanimous checking matrix.