基于大数逻辑可译LDPC码的译码算法研究

Decoding Algorithms for Majority-Logic Decodable LDPC Codes

本文提出两种基于可靠度的迭代大数逻辑译码算法,从以下两个方面降低译码复杂度：（1）校验节点使用伴随式信息处理,可节省外信息的计算操作;（2）变量节点使用伴随信息进行总信息的投票计数过程.结合非均匀量化技术,接收信号在判决门限附近获得更加精细的处理.此外,本文利用量化参数和列重比例信息对可靠度偏移方向和幅度进行了设计.仿真实验表明,本文提出的算法能够在很低的量化比特（3～4 bits）下有效工作,具有优良的译码性能和快速的收敛速度.

In this paper,we present two iterative reliability-based majority-logic decoding algorithms. The proposed algorithm has a lower complexity due to the following two modifications. First,at check nodes,syndrome messages instead of extrinsic messages are calculated and passed back to variable nodes. Second,at the variable nodes,the syndrome messages are directly involved in the voting process. The non-uniform quantization scheme and a newiterative process of reliability-message shifting are also presented in this paper. The received signal achieve a higher quantization resolution near the decision threshold. The shifting direction / step are jointly designed with the quantization parameter and the column weights. Simulation results showthat,the presented algorithms can achieve excellent BER performance and fast decoding speed,even with very small quantization levels（ 3 ～4 bits resolution）.