一种基于比特翻转的极化码FSCLF译码算法

A FSCLF Decoding Algorithm of Polar Codes Based on Bit-Flipping

摘要:

针对串行抵消列表翻转(Successive Cancellation List Flip,SCLF)译码算法存在译码性能与复杂度不能同时兼顾的问题,提出了一种快速串行抵消列表翻转(Fast Successive Cancellation List Flip,FSCLF)译码算法。该算法通过加入四种特殊结点的识别来加快译码速率,同时构建了临界集(Critical Set,CS),不再依据先前译码错误而引起的错误传播,而是通过两种特殊结点即信息比特R1结点和单奇偶校验(Single-Parity-Check,SPC)结点分别对对数似然比(Log-Likelihood Ratio,LLR)值进行计算来判决并确定翻转位置,当奇偶校验位不满足时只需翻转与最不可靠输入LLR值相对应的信息比特,这样减少了翻转次数,从而降低了算法复杂度。仿真结果表明:在误块率为10-5时,所提出的FSCLF译码算法比原SCLF译码算法的信噪比改善了0.09dB,为中短码长情况提供了参考算法。

Abstract:

A fast successive cancellation list flip (FSCLF) decoding algorithm is proposed to improve the problem that the decoding performance and complexity of the existing successive cancellation list flip (SCLF) decoding algorithm cannot be considered at the same time. The proposed algorithm speeds up the decoding by adding the identification of four special nodes. In the meantime, the construction of the critical set (CS) no longer depends on the error propagation caused by the previous decoding errors, instead, the log-likelihood ratio (LLR) value is calculated to determine the flipping position by the two special nodes of the R1 node and single-parity-check (SPC) node, and the information bit corresponding to the least reliable input LLR only need to be flipped when the parity check bit isn’t satisfied, thus both the number of the flipping and the complexity is reduced. Simulation results show that the signal-to-noise ratio of the FSCLF algorithm, compared with the SCLF algorithm, is improved about 0.09dB at the block error ratio (BLER) of 10-5. Therefore, the proposed decoding algorithm can play an important role in the middle and short code-length.