该文从译码速率、硬件实现的复杂度和误码率3个方面对比研究了两种典型的高速译码算法:Turbo型和积算法与并行加权比特翻转算法。以准循环LDPC码为对象,给出了Turbo型和积算法和并行加权比特翻转算法的实现时序、硬件复杂度以及误码率性能,其中,并行加权比特翻转算法的高效时序结构是首次给出的。计算机仿真结果表明,这两种算法都能够在迭代次数较少时取得良好的性能。
In this paper, two typical high throughput decoding algorithms for QC-LDPC codes are presented: turbo-min-sum-product algorithm and parallel-weighted-bit-flipping algorithm. Their performance are compared in three aspects: decoding speed, hardware complexity, and error rate. In order to achieve high-throughput, a novel time scheduling for parallel-weighted~bit-flipping algorithm is proposed. Computer simulation confirms the effectiveness of two algorithms.