基于LDPC编码器的多码率并行算法的设计研究
Design Study of Multi-Code Rate Parallel Algorithm Based on LDPC Encoder
孙紫檀 1周益 1易小谦 1赵臣 1李生贵1
作者信息
- 1. 中国商用飞机有限责任公司民用飞机试飞中心,上海 201323
- 折叠
摘要
为了提高通信信息传输过程中的译码速度,研究以低密度校验码为对象,选取了置信传播译码算法为研究对象,并对其进行并行性能优化.该优化过程分为两步,第一步是采用对数似然比法对算法的译码过程进行运算简化,第二步是通过通用处理器提升算法的并行速率.对优化算法的性能进行验证后,结果表明置信传播译码算法在信噪比为4.0时,BP算法的耗时为372 ms.优化后的置信传播译码算法耗时降低到了 287 ms.故研究提出的优化算法可用于通信领域的并行译码工作中.
Abstract
Low-density parity check codes were investigated in order to increase the decoding speed during communication infor-mation transfer.The confidence propagation decoding algorithm was chosen as the research subject,and its parallel performance was optimized.There are two steps in the optimization process.The first stage is to use the logarithmic likelihood ratio approach to simplify the algorithm's decoding process.The second step is to use a universal processor to increase the algorithm's parallel speed.The con-fidence propagation decoding process takes 372 ms when the signal-to-noise ratio is 4.0,according to the results of the optimization algorithm's performance verification.The time consumption drops to 287 ms with the enhanced confidence propagation decoding tech-nique.Therefore,the optimization algorithm proposed in the study can be used in parallel decoding work in the field of communica-tion.
关键词
LDPC/多码率/并行运算/BP算法/GPUKey words
LDPC/multiple code rates/parallel algorithm/BP algorithm/GPU引用本文复制引用
出版年
2024
NETL
NSTL
万方数据