低复杂度的自适应置信差分迭代译码算法

An Adaptive Belief Propagation Difference-map Iterative Decoding Algorithm with Low Complexity

针对中短码长的低密度奇偶校验规则码(Low Density Parity Check,LDPC)规则码,该文采用消息更新规则改进和因子图变换方法,提出一种低复杂度差分迭代译码算法。在置信传播算法的基础上,仅当变量节点的消息值振荡时引入差分映射策略,得出一种选择性的置信差分规则,自适应地调整校验节点消息的归一化系数,提高译码性能。同时,采用展开校验节点的图变换方法,将计算复杂度从随节点度分布指数性增长降至线性增长。分别在高斯白噪声信道和瑞利衰落信道下进行仿真实验,结果表明该算法和基于图变换的其他低复杂度译码算法相比,性能优越且复杂度低,和对数似然比的置信传播算法(LLR-BP)相比,高信噪比区域内的性能优异,低信噪比区域内的计算复杂度明显降低。

In this paper, an adaptive belief difference-map propagation algorithm with low complexity is proposed for short and middle length LDPC regular codes by modifying message update rules and transforming factor graph. To improve decoding performance, a new selective belief propagation difference-map message update rule is introduced by borrowing the difference-map strategy for variable node messages oscillation, and the normalized factor is adjusted adaptively. Meanwhile, the computational complexity exponential in the degree of check node is decreased into linear in degree by opening the check node. The simulation results illustrate that the proposed algorithm has better performance and lower complexity than other iterative decoding algorithms based on the modified factor graphs. Compared to the LLR-BP, it better performance at high Eb/N0 and the computational complexity is apparently downgraded at low Eb/N0.