Two classes of QC-LDPC cycle codes approaching Gallager lower bound
Two classes of QC-LDPC cycle codes approaching Gallager lower bound
复制标题
DOI:
10.1007/s11432-018-9778-x
复制
发表时间:
2019-06
期刊:
影响因子:
--
通讯作者:
Hengzhou Xu;Huaan Li;Mengmeng Xu;Dan Feng;Hai Zhu
中科院分区:
文献类型:
--
作者:
Hengzhou Xu;Huaan Li;Mengmeng Xu;Dan Feng;Hai Zhu
Dear editor, Many research results show that, for equivalent bit length, short nonbinary LDPC codes outperform binary LDPC codes by about 1 dB [1–3]. Moreover, nonbinary LDPC codes have lower errorfloor, fast iterative decoding convergence, and strong ability of correcting burst errors. But the roadblock to their application is the high decoding complexity. Recently, significant studies on the low-complexity decoding algorithms of nonbinary LDPC codes have been done [4]. It is noticeable that these low-complexity algorithms are under the frame of iterative decoding. Hence, it is interesting to design nonbinary LDPC codes with large minimum distance and suitable for the iterative algorithms.For a given block length, nonbinary LDPC codes perform better and better with the increase of their finite field size. When the finite field size is sufficiently large, the increased coding gain becomes negligible, and then the column weight in the parity-check matrices of the best nonbinary LDPC codes tends to 2. In order to facilitate the hardware implementation, quasi-cyclic (QC) structure should be considered. In this study, we study a class of binary QC-LDPC codes with column weight 2 and row weight ρ. Notice that this class of codes is referred to as (2, ρ)-regular QC-LDPC cycle codes. By replacing 1’s in the parity-check matrices of QC-LDPC cycle codes with nonzero elements of nonbinary finite fields,