The Weighted Information Storage Bit-Flipping Decoder for Generalized LDPC Codes
Haobin Xu, Zhongfeng Wang, Suwen Song
By incorporating stronger component codes, generalized Low-Density-Parity-Check (GLDPC) codes significantly enhance error-correction capability. Bit-flipping (BF) algorithms are among the primary hard-decision decoding algorithms for GLDPC codes. However, existing BF algorithms for GLDPC still exhibit limited decoding performance. This paper proposes a Weighted Information Storage Bit-Flipping (WISBF) decoding algorithm for GLDPC codes with different types of component codes to improve the decoding performance, which introduces a two-stage energy-update strategy to reduce the percentage of erroneous flips and lower the average number of iterations. The simulation results demonstrate that the WISBF algorithm far outperforms the state-of-the-art Weighted Bit-Flip Voting (WBFV) on the tested GLDPC code, achieving an approximately three-order-of-magnitude improvement in frame-errorrate (FER). Moreover, the GLDPC codes with WISBF also achieve superior performance over their LDPC counterparts decoded using existing bit-flipping (BF) algorithms, including the Information Storage Bit-Flipping (ISBF) algorithm. We then present a detailed hardware architecture for the WISBF decoder. Synthesized under the TSMC $90-\mathrm{nm}$ technology, the proposed GLDPC WISBF decoder achieves $\mathbf{1 1 4. 3 \%}$ higher throughput than the GLDPC WBFV decoder.
IEEE International Conference on Electronics, Communications and Intelligent Science (ECIS)•May 22, 2026•0 citations