Adaptive Group Based Symbol Flipping Decoding Algorithm.

This paper introduces an adaptive technique for grouping variable nodes in decoding non-binary LDPC codes that involve symbol flipping. The technique considers both the individual symbol reliability and majority-based voting when grouping variable nodes in each iteration. Groups are formed based on the cumulative density of the variable nodes, with the least reliable variable nodes given the highest priority. The decoding algorithm then proceeds to decode subsequent groups in order of decreasing priority. The results of numerical analysis demonstrate that this approach strikes a balance between computational complexity and bit error rate performance, making it suitable for various applications, such as data storage and mobile networks.