一种非规则广义LDPC码的构造方法

A Construction Method of Irregular Generalized LDPC Codes

广义低密度奇偶校验(Generalized Low-Density Parity-Check, GLDPC)码把低密度奇偶校验(Low-Density Parity-Check, LDPC)码中的单奇偶校验(Single Parity-Check, SPC)节点替换为校验能力更强的广义约束(Generalized Constraint, GC)节点,使其在中短码和低码率的条件下具有更低的误码率。传统GLDPC码要求基矩阵的行重等于分量码的码长,这限制了GLDPC码构造的灵活性。另外,相比于传统GLDPC码中GC节点位置的随机选取,GC节点的位置选择在GLDPC码的误码率性能上有一定的优化空间。针对以上两点,提出了一种基于渐进边增长(Progressive Edge-Growth, PEG)算法的非规则GLDPC码构造方法和一种基于Tanner图边数的GC节点位置选择算法。使用PEG算法生成的非规则LDPC码作为本地码,根据本地码的校验节点度使用多种分量码,结合GC节点位置选择算法构造非规则GLDPC码。仿真结果表明,与传统方法构造的GLDPC码相比,基于Tanner图边数的GC节点位置选择算法构造的非规则PEG-GLDPC码在误码率和译码复杂度上均得到明显改善。

Generalized low-density parity-check(GLDPC)codes replace the single parity-check(SPC)nodes in low-density parity-check(LDPC)codes with more powerful generalized constraint(GC)nodes,resulting in a lower bit error rate(BER)at short to medium codes and low code rates.Conventional GLDPC code requires the row weights of the base matrix to be equal to the code length of the component code,which limits the flexibility of GLDPC code construction.In addition,compared with the random selection of GC nodes location in conventional GLDPC code,the location selection of GC nodes has room for improvement in the BER performance of GLDPC code.For above two problems,a construction method of irregular GLDPC code based on the progressive edge-growth(PEG)algorithm and a GC nodes location selection algorithm based on the number of the edges in Tanner graph are proposed.The irregular LDPC code generated by the PEG algorithm is used as local code.And then,according to the check node degree of the local code,multiple component code is used.Finally,the irregular GLDPC code is constructed by combining the GC nodes location selection algorithm.The simulation results show that,the irregular PEG-GLDPC code constructed by combined with the GC nodes location selection algorithm based on the number of the edges in Tanner graph is significantly improved in terms of BER and decoding complexity compared with the GLDPC codes constructed by the conventional method.