基于信息截断的低复杂度多进制LDPC码译码器

Low-complexity Decoder for Non-binary LDPC Codes Based on Truncated Information

  • 摘要: 针对多进制低密度奇偶校验(LDPC)码译码算法实现复杂度较高的问题,基于简化增强串行广义比特翻转译码算法(SES-GBFDA),提出将每个符号的对数似然值截断为有限值进而有效减少存储需求和计算复杂度的译码算法,即截断SES-GBFDA。对于定义在伽罗华域GF(32)上的多进制LDPC码译码器,将基本更新单元的数量由32个减少为10个来完成变量节点消息的处理,显著降低了计算复杂度。在现场可编程门阵列(FPGA)上实现了定义在GF(32)上,码长为837个符号,码率为0.85的多进制LDPC码译码器。结果表明,译码器的吞吐量可以达到90 Mbps,与未进行对数似然值截断的译码器相比,所实现译码器在译码性能损失0.25 dB的情况下,将查找表和寄存器资源消耗分别减少了64.5%和76.3%。

     

    Abstract: For the high implementation complexity of non-binary low-density parity-check (LDPC) codes decoding algorithm, based on the simplified enhanced serial generalized bit flipping decoding algorithm (SES-GBFDA), a method truncating the log-likelihood value for each symbol to a finite value is proposed, which can effectively reduce the memory requirement and the computational complexity, that is, truncated SES-GBFDA. For the non-binary LDPC decoder defined on the Galois Field GF(32), the number of basic update units was reduced from 32 to 10 to complete the processing of variable node messages, thereby significantly reducing the computational complexity. A non-binary LDPC decoder defined over GF(32) with a code length of 837 symbols and a code rate of 0.85 was implemented on the field programmable gate array (FPGA). The results show that, the throughput of the decoder can reach 90 Mbps, compared with the decoder without log-likelihood value truncation, the implemented decoder reduces the consumption of look-up table and register resources by 64.5% and 76.3%, respectively, with 0.25 dB loss in decoding performance.

     

/

返回文章
返回