极化码的置信传播译码算法优化

Optimized Belief Propagation Decoding Algorithm for Polar Codes

极化码是信道编码的里程碑成果。串行抵消列表(SCL)译码取得较好的误码性能,但译码时延较大。置信传播(BP)算法可以有效减少译码时延和计算复杂度,然而现有的BP算法的译码因子图存在大量短环,其译码性能远远不及SCL译码算法。本文在详尽分析现有的BP译码算法后,提出了基于置换因子图的置信传播列表(CA-BPL)译码算法。仿真表明,该算法可以有效提高极化码的译码性能。

Polar codes are a milestone in channel encoding.The successive cancellation list(SCL)decoding algorithm achieves good error performance,it introduces a long decoding delay.The belief propagation(BP)algorithm can avoid large delay.However,the factor graph of the polar codes has a large number of short loops,which make decoding performance much poorer than the SCL algorithm.After analysis of the existing BP decoding algorithm,we propose a belief propagation list(CA-BPL)decoding algorithm based on random permutation factor graph.The simulation show that the scheme can improve the decoding performance of BP algorithm.