BICM-ID系统中的均匀重加权迭代译码算法

针对LDPC编码的BICM-ID系统,建立了正规因子图模型。从消息更新规则的角度,运用变分方法统一解释了均匀重加权置信传播算法和标准置信传播算法。将均匀重加权方法提高性能的特点与置信传播的特性相结合,推导出均匀重加权迭代译码算法,讨论了边出现概率对算法性能的影响。分别在高斯白噪声信道和不同调制方式下进行仿真实验。结果表明,提出的均匀重加权迭代译码算法性能均优于标准置信传播和差分映射置信传播等迭代译码算法。

Uniformly reweighted iterative decoding algorithm in BICM-ID system

The LDPC coded bit-interleaved coded modulation iterative decoding (BICM-ID) system was modeled by a normal factor graph. From the angle of the message update rule，uniformly reweighted belief propagation algorithm (URW-BP) and belief propagation algorithm (BP) are unified and interpreted as message passing algorithm by variation method. Combining the local nature of BP with the improved performance of URW-BP in graphs with cycles, the uniformly reweighted iterative decoding algorithm (URID) was derived and the impact of edge appearance probabilities (EAP) on its performance discussed. Simulation results for AWGN and Rayleigh fading channel and different modulation illustrate the performance improvement of URID compared with other iterative decoding algorithms such as BP and Difference-Map BP（DMBP）algorithm.