论文部分内容阅读
在采用基于对数似然比的置信度传递译码算法LLR-BP(Log-Likelihood Ratio Belief Propagation)对低密度奇偶校验码LDPC code(Low-density Parity-check code)译码时,变量节点对数似然比外信息ex-LLR(ExtrinsicLog-Likelihood Ratio)的震荡会引起比特节点发生错误。文中分析了ex-LLR值和后验概率值(APP)产生震荡的原因及两者的关系,提出了减弱其对整个译码过程影响的两种改进译码算法。仿真结果表明,改进的译码算法在基本不增加译码复杂度的情况下,能够降低译码错误平层,进一步改善传统的LLR-BP译码算法的性能。
In the decoding of low-density parity-check codes (LDPC codes) using Log-Likelihood Ratio Belief Propagation (LLR-BP) based on log-likelihood ratio, variable nodes Oscillation of log-likelihood ratio ex-LLR (ExtrinsicLog-Likelihood Ratio) can cause bit node errors. In this paper, we analyze the causes of the ex-LLR value and the posterior probability value (APP) and their relationship, and propose two improved decoding algorithms to reduce their impact on the entire decoding process. The simulation results show that the improved decoding algorithm can reduce the decoding error level without any increase of decoding complexity and further improve the performance of the traditional LLR-BP decoding algorithm.