削弱震荡节点的LDPC码译码算法研究

摘要

为了进一步提高动态调度算法的误码性能,提出一种基于变量节点消息震荡的残差置信传播(Residual Belief Propagation based on Variable Nodes message Oscillation phenomenon,VNO-RBP)算法.该算法以面向节点的残差置信传播(Node-Wise RBP,NW-RBP)算法为基础,当译码器出现震荡后,以变量节点消息LLR(Log Likelihood Ratio)值的相对可靠度作为参考,选择相对残差值较大的节点优先更新,加快译码收敛速度.同时,选取译码过程中产生的震荡节点作为目标节点,将目标节点前后两次迭代中的LLR值作加权平均处理,降低该节点的不可靠度.理论分析及仿真结果表明,本文提出的VNO-RBP译码算法比NW-RBP算法复杂度更低,在误码率为10-5时误码性能提升大约0.31 dB.%In order to improve the BER performance of dynamic scheduling algorithm,the VNO-RBP (based on VN message oscillation of RBP) algorithm was proposed in this paper.This algorithm was based on the NW-RBP algorithm,when the decoder began to oscillate,we would select the nodes with larger residual that need prior updating by making reference to the relative reliability of variable nodes in the process of decoding so as to speed the decoding speed.At the same time,selecting the target nodes that arise oscillation,executing weighted mean processing to LLR values generated before and after updating so as to reduce the unreliability of variable nodes.The theoretical analysis and simulation results demonstrate that compared with NW-RBP algorithm,the proposed VNO-RBP decoding algorithm has lower decoding complexity and its BER performance improves 0.31 dB when the error rate reaches 10-5.