摘要
近年来,基于置信传播(BP),最小和(MS)和归一化最小和(NMS)算法,已经提出3种相对应的LDPC码加权比特翻转(WBF)译码算法。但这3种WBF算法所代表的物理意义和内在的紧密联系问题目前仍未有所研究。该文依据一种全新的理解方式,对3种WBF算法进行理论推导,并阐述3种算法内在的紧密联系,最后通过仿真验证所得结论的合理性和正确性。这对于设计新的改进型WBF算法具有一定的指导意义。
Recently, based on Belief-Propagation (BP), Min-Sum (MS) and Normalized MS (NMS) algorithms, three corresponding Weighted Bit Flipping (WBF) decoding algorithms are proposed for LDPC codes. However, not only the strict physical significance but also the inherent relationship of these WBF algorithms is still remain largely unknown. In this paper, the theoretical derivation, and an inherent relationship between them is developed from a whole novel understanding. Furthermore, the simulation results demonstrate the rationality and accuracy of the conclusion, which presents a certain reference value for design of new improved WBF algorithms.
出处
《电子与信息学报》
EI
CSCD
北大核心
2014年第9期2093-2097,共5页
Journal of Electronics & Information Technology
基金
国家自然科学基金(61032003
61271172)
中央高校基本科研业务费专项基金(A03008023901004)
博士点基金(20120185110030
2013018530002)资助课题
关键词
低密度奇偶校验码
加权比特翻转
可靠度后验信息
对数最大后验概率
Low-Density Parity-Check (LDPC) codes
Weighted Bit Flipping (WBF)
Posterior reliabilityinformation
Log maximum a posteriori