摘要
置信传播(beliefpropagation,BP)算法的计算复杂度较高,且变量、校验节点间信息传递时的振荡现象导致了译码性能降低。为了提高其译码性能、降低该算法的计算复杂度,通过分析校验和与错误位数的关系,提出了一种提前结束迭代的判决准则;另外,通过对振荡信息的修正来改善其传递的信息的可靠性,进而来提高BP算法的译码性能。仿真结果显示,改进的译码算法在降低了低信噪比时的平均迭代次数的同时还提高了译码性能。
The computational complexity of BP algorithm is high,and the oscillation is observed when the information passed between variable nodes and check nodes,which leads to decoding performance degraded.In order to improve its decoding performance and reduce the computational complexity,by analyzing the relationship between checksum and the number of error bits,an method of early stopping criterion is proposed to reduce the number of iteration.In addition,through correction of the oscillation information to improve the reliability of the information transmission,and then to improve decoding performance of BP algorithm.Simulation results show that with the new decoding algorithm,the average number of iteration is reduced when SNR is low and the decoding performance is improved.
出处
《科技通报》
北大核心
2013年第3期111-115,共5页
Bulletin of Science and Technology
基金
国家自然科学基金委员会与中国工程物理研究院联合基金资助项目(10776007)