A new decoding strategy of belief propagation (BP) for low‐density parity‐check codes is presented by serially concatenating the traditional simultaneous decoding (also known as flooding) and the informed dynamic decoding (IDD) (e.g. node‐wise residual belief propagation (NW RBP) or informed variable‐to‐check (IVC) RBP). The frame error rates (FER) and bit error rates (BER) of the concatenated BP decoder outperform that of the non‐concatenated single decoder. In addition, as the signal‐to‐noise ratio (SNR) increases, the average number of decoding iterations required for the concatenated BP decoder decreases, and tends to merge with that of the first sub‐decoder in the high SNR region. Moreover, since the IDD strategies concerned in this paper require extra effort for the residual computation, and the sequential updating of IDD is much more time consuming than the simultaneous updating of flooding. Therefore, compared with using a single IDD, serially concatenating flooding and IDD reduces both the approximate decoding complexity and the decoding latency because the concatenated decoder is still dominated by the sub‐decoder flooding.

In this paper, a novel extending algorithm is proposed for the construction of rate compatible (RC) irregular repeat accumulate like (IRA-like) codes. Based on the structure characters of IRA-like codes, we propose an extending algorithm with uniform check node degree distribution and strong dependency in parity check matrix. Lower rate codes constructed through this method are efficiently encodable and well performed. Starting from initial code rate 8/16, the designed extending algorithm combining with existing puncturing algorithm realizes a series of RC IRA-like codes with rate R = 8/24, 8/23... 8/9. Simulation results show that the constructed RC IRA-like codes outperform those of similar algorithms from 0.04dB to 0.4 dB.

A multiple-vote symbol-flipping (MV-SF) decoding algorithm for nonbinary low-density parity-check (NB-LDPC) codes is proposed in this paper. Our algorithm improves the generalized bit-flipping algorithm (GBFDA) by considering the multiplicity of the candidates at the check-node output, to perform a more accurate symbol-flipping decision at the variable node update. The MV-SF algorithm greatly improves the frame error rate performance of GBFDA and approaches the performance of the best state-of-the-art decoders [extended min-sum and min-max (Min–Max)] with lower complexity. For a$(N=837,K=723)$NB-LDPC code over GF(32), the decoder derived from the proposed algorithm can reach a throughput higher than 500 Mb/s and a coding gain of 0.44 dB compared with the most efficient GBFDA architecture with only twice the silicon area. Our architecture has 27% efficiency gain compared with the best Min–Max architecture found in the literature, with a performance loss of just 0.21 dB at frame error rate$10^{-4}$.

Recently, we introduced a new class of finite alphabet iterative decoders (FAIDs) for low-density parity-check (LDPC) codes. These decoders are capable of surpassing belief propagation (BP) in the error floor region on the binary symmetric channel (BSC) with much lower complexity. In this paper, we introduce a novel scheme with the objective of guaranteeing the correction of a given and potentially large number of errors on column-weight-three LDPC codes. The proposed scheme uses a plurality of FAIDs which collectively correct more error patterns than a single FAID on a given code. The collection of FAIDs utilized by the scheme is judiciously chosen to ensure that individual decoders have different decoding dynamics and correct different error patterns. Consequently, they can collectively correct a diverse set of error patterns, which is referred to as decoder diversity. We provide a systematic method to generate the set of FAIDs for decoder diversity on a given code based on the knowledge of the most harmful trapping sets present in the code. Using the well-known column-weight-three (155,64) Tanner code with dmin= 20 as an example, we describe the method in detail and show, by means of exhaustive simulation, that the guaranteed error correction capability on short length LDPC codes can be significantly increased with decoder diversity.

In this paper, we present an improvement and a new implementation of a simplified decoding algorithm for non-binary low density parity-check codes (NB-LDPC) in Galois fields GF(q). The base algorithm that we use is the Extended Min-Sum (EMS) algorithm, which has been widely studied in the recent literature, and has been shown to approach the performance of the belief propagation (BP) algorithm, with limited complexity. In our work, we propose a new way to compute modified configuration sets, using a trellis representation of incoming messages to check nodes. We call our modification of the EMS algorithm trellis-EMS (T-EMS). In the T-EMS, the algorithm operates directly on the deviation space by considering a trellis built from differential messages, which serves as a new reliability measure to sort the configurations. We show that this new trellis representation reduces the computational complexity, without any performance degradation. In addition, we show that our modifications of the algorithm allows to greatly reduce the decoding latency, by using a larger degree of hardware parallelization.

We introduce a generic approach, called FAID diversity, for improving the error correction capability of regular low-density parity check codes, beyond the belief propagation performance. The method relies on operating a set of finite alphabet iterative decoders (FAID). The message-passing update rules are interpreted as discrete dynamical systems, and are judiciously chosen to ensure that decoders have different dynamics on a specific finite-length code. An algorithm is proposed which uses random jumps in the iterative message passing trajectories, such that the system is not trapped in periodic attractors. We show by simulations that the FAID diversity approach with random jumps has the potential of approaching the performance of maximum-likelihood decoding for finite-length regular, column-weight three codes.