分层近似规则LDPC码的编码器设计

提出一种分层近似规则(LAR)LDPC码的构造方法及其编码器的设计方案.该方案在现有的RU算法的基础上,完全去掉了前向替换(FS)的步骤,并引入循环移位寄存器结构来处理密矩阵与向量的乘法,使其硬件复杂度从与密矩阵维数平方成正比,下降到只与其维数成正比.与RU算法相比,新方案缩短了编码器的编码延时,提高了吞吐量,还对不同码长和码率的应用具有线上重构的灵活性.仿真结果表明,分层近似规则LDPC码具有与随机构造的规则码极其相近的纠错性能,具有很高的实用参考价值.

An Encoder Design for Layered Approx-Regular LDPC Codes

A construction method and an encoder design scheme for layered approxregular low-density-parity-check(LAR-LDPC) codes were proposed.Although based on the existing Richardson-Urbanke(RU) algorithm,the new scheme entirely removes the forward substitution(FS) operation and handles the dense-matrix-vector multiplication using cyclic shift registers.This decreases the hardware complexity from linearity to the square of the dimension of dense matrix to linearity only to the dimension of dense matrix.Compared with RU algorithm,the new scheme reduces the encoding latency and improves the throughput,and features high flexibility of online reconfiguration in both code length and code rate.The simulation result shows that LAR-LDPC codes closely approach randomly constructed regular LDPC codes in terms of error correction performance,thus enjoying a fairly high referential merit in practical application.